Wind power converter speed sensorless field-oriented control method and system

Abstract

The application relates to the technical field of intelligent control, and provides a wind power converter speed sensor-free magnetic field oriented control method and system, which introduces an accurate rotor resistance online hybrid identification mechanism, estimates the rotor resistance and the time constant in real time, and deeply integrates the rotor resistance and the time constant into the flux linkage estimation and the speed self-adaptive process, so that the contradiction between parameter identification and speed estimation is effectively decoupled. In this way, when the wind power converter faces complex environment temperature fluctuations and low-speed working conditions, the control precision and stability of the wind power converter are significantly improved, the flux linkage estimation deviation and system out-of-control are effectively avoided, and the speed sensor-free control technology is more robust and reliable.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to a sensorless magnetic field orientation control method and system for wind power converters. Background Technology

[0002] In wind power generation systems, the wind converter is one of the core components, and its efficient and reliable operation is crucial to the performance of the entire power generation system. Traditional wind converter field-oriented control schemes typically rely on mechanical speed sensors to obtain rotor speed and position information. However, these sensors not only increase the hardware cost and complexity of the system but also make it vulnerable to damage in the harsh operating environment of wind farms, thereby reducing the overall reliability of the system and increasing the maintenance burden. Therefore, developing a sensorless field-oriented control method that estimates rotor speed and position using purely electrical quantities has significant engineering implications for improving the economy, robustness, and service life of wind converters.

[0003] In existing technologies, sensorless field-oriented control schemes have been widely researched and applied. Among them, observers based on Model Reference Adaptive Systems (MRAS) have attracted much attention due to their simple structure and ease of implementation. However, these methods still face significant technical challenges in practical wind power applications. Particularly noteworthy is that when rotor resistance fluctuates drastically due to temperature changes, the rotor time constant, which determines the dynamic response of the flux linkage, also changes significantly. In this case, if the rotor time constant used in the controller does not match the actual value of the motor, even if the speed estimation is accurate, the flux linkage estimation will be inaccurate. In the low-speed operating region of wind power converters, the back electromotive force component of the motor is small, while the influence of rotor time constant changes on flux linkage calculation is dominant. This makes it easy for the original adaptive mechanism to incorrectly attribute all flux linkage errors to inaccurate speed estimation, leading to positive feedback in the system, outputting completely incorrect speed and magnetic field information, and even causing system runaway. This decoupling error caused by model simplification has become the main bottleneck for existing sensorless field-oriented control methods to achieve high-precision and high-reliability control over a wide temperature range. Summary of the Invention

[0004] This invention provides a sensorless field-oriented control method and system for wind power converters. It introduces a precise online hybrid identification mechanism for rotor resistance to estimate rotor resistance and time constant in real time, deeply integrating these parameters into the flux estimation and speed adaptation processes. This effectively decouples the contradiction between parameter identification and speed estimation. In this way, the control accuracy and stability of the wind power converter are significantly improved when facing complex environmental temperature fluctuations and low-speed conditions, effectively avoiding flux estimation errors and system runaway, making sensorless control technology more robust and reliable.

[0005] In a first aspect, the present invention provides a sensorless field-oriented control method for a wind power converter, comprising: The three-phase stator current and three-phase stator voltage of the wind power converter are transformed by coordinate to obtain the dq-axis stator current and dq-axis stator voltage; The flux linkage is estimated based on the voltage model for the dq-axis stator current and dq-axis stator voltage to obtain the reference rotor flux linkage; Online hybrid identification of rotor resistance is performed on the dq-axis stator current and dq-axis stator voltage to obtain estimated rotor resistance and estimated rotor time constant; Based on the estimated rotor time constant, flux linkage estimation and speed adaptation based on the current model are performed on the dq-axis stator current to obtain the estimated rotor mechanical angular velocity and the estimated rotor magnetic field position. The estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command are used to perform field-oriented control to obtain the PWM gate drive signal.

[0006] Secondly, the present invention provides a sensorless magnetic field orientation control system for a wind power converter, comprising: The coordinate transformation module is used to perform coordinate transformation on the three-phase stator current and three-phase stator voltage corresponding to the wind power converter to obtain the dq-axis stator current and dq-axis stator voltage; The flux linkage estimation module is used to perform voltage model-based flux linkage estimation on the dq-axis stator current and dq-axis stator voltage to obtain the reference rotor flux linkage. The online hybrid identification module for rotor resistance is used to perform online hybrid identification of rotor resistance based on the dq-axis stator current and dq-axis stator voltage to obtain an estimated rotor resistance and an estimated rotor time constant. The speed adaptive module is used to perform flux linkage estimation and speed adaptation based on the current model of the dq-axis stator current based on the estimated rotor time constant, so as to obtain the estimated rotor mechanical angular velocity and the estimated rotor magnetic field position. The field orientation control module is used to perform field orientation control on the estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command to obtain the PWM gate drive signal.

[0007] Compared with existing technologies, this invention provides a sensorless field-oriented control method and system for wind power converters. By introducing a precise online hybrid identification mechanism for rotor resistance, it estimates rotor resistance and time constant in real time and deeply integrates them into the flux estimation and speed adaptation processes, effectively decoupling the contradiction between parameter identification and speed estimation. In this way, the control accuracy and stability of the wind power converter are significantly improved when facing complex environmental temperature fluctuations and low-speed conditions, effectively avoiding flux estimation errors and system runaway, making the sensorless control technology more robust and reliable. Attached Figure Description

[0008] One or more embodiments are illustrated by way of example with the corresponding pictures in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0009] Figure 1 A flowchart of a sensorless magnetic field orientation control method for a wind power converter according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the data flow in a sensorless magnetic field orientation control method for a wind power converter according to an embodiment of the present invention. Figure 3 This is a block diagram of a sensorless magnetic field orientation control system for a wind power converter according to an embodiment of the present invention. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the various embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and with various variations and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0011] The present invention proposes a sensorless magnetic field orientation control method for wind power converters. Figure 1 This is a flowchart of a sensorless magnetic field orientation control method for a wind power converter according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the data flow in a sensorless field-oriented control method for a wind power converter according to an embodiment of the present invention. (In conjunction with...) Figure 1 and Figure 2According to an embodiment of the present invention, a sensorless field-oriented control method for a wind power converter includes the following steps: S1, performing coordinate transformation on the three-phase stator current and three-phase stator voltage of the wind power converter to obtain the dq-axis stator current and dq-axis stator voltage; S2, performing flux linkage estimation based on a voltage model on the dq-axis stator current and dq-axis stator voltage to obtain a reference rotor flux linkage; S3, performing online hybrid identification of rotor resistance on the dq-axis stator current and dq-axis stator voltage to obtain an estimated rotor resistance and an estimated rotor time constant; S4, based on the estimated rotor time constant, performing flux linkage estimation and speed adaptation based on a current model on the dq-axis stator current to obtain an estimated rotor mechanical angular velocity and an estimated rotor magnetic field position; S5, performing field-oriented control on the estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command to obtain a PWM gate drive signal.

[0012] Specifically, S1 involves performing a coordinate transformation on the three-phase stator current and three-phase stator voltage of the wind power converter to obtain the dq-axis stator current and dq-axis stator voltage. It should be understood that the stator current and voltage of a three-phase AC motor (such as a wind turbine) are essentially AC quantities that vary sinusoidally with time. Directly and precisely controlling these time-varying AC quantities, especially achieving independent decoupling control of torque and flux linkage, is extremely complex in terms of algorithm design. To simplify controller design and improve control performance, these physical quantities, which appear as AC waveforms in a stationary three-phase coordinate system (abc coordinate system), need to be transformed into a coordinate system that rotates synchronously with the rotor magnetic field (dq coordinate system). In the synchronously rotating coordinate system, these physical quantities will behave as DC quantities during steady-state operation. Therefore, by drawing on the control concepts of DC motors and employing mature proportional-integral (PI) controllers and other technologies, precise and rapid adjustment can be achieved, laying the foundation for subsequent field-oriented control.

[0013] In its specific implementation, S1 first performs a Clarke transformation to convert the stator current and stator voltage in the three-phase stationary coordinate system (abc) to the two-phase stationary coordinate system (α-β), obtaining the corresponding α-β axis components. This transformation process converts three related AC quantities into two mutually orthogonal AC quantities.

[0014] Next, the Park transformation is performed to rotate the physical quantities in the two-phase stationary coordinate system (α-β) to the two-phase rotating coordinate system (dq), thereby obtaining the final required dq-axis stator current and dq-axis stator voltage.

[0015] For the transformation of three-phase stator voltage, by replacing the three-phase stator current in the Clark transformation and Park transformation with the three-phase stator voltage, the dq-axis stator voltage, including the d-axis stator voltage and the q-axis stator voltage, can be obtained.

[0016] Specifically, in S2, flux linkage estimation based on a voltage model is performed on the dq-axis stator current and dq-axis stator voltage to obtain a reference rotor flux linkage. It should be understood that in field-oriented control theory, rotor flux linkage is a core state variable, and its accurate observation is a prerequisite for achieving decoupled torque and flux linkage control. The voltage model is a classic flux linkage observation method, which mainly calculates the flux linkage based on the stator voltage equation of the motor. Its significant advantage is its insensitivity to changes in motor parameters (especially rotor resistance, which is susceptible to temperature effects) in the medium-to-high speed range, thus providing a relatively stable and accurate flux linkage estimate. In the technical solution of this invention, the rotor flux linkage calculated by the voltage model is not directly used in the speed feedback loop, but rather serves as a reference term, i.e., the ideal flux linkage size that the control system expects to maintain. This reliable reference value provides an accurate alignment target for subsequent flux linkage estimation and speed adaptive circuitry based on the current model, effectively avoiding estimation drift and system runaway problems that may occur due to parameter mismatch when using the current model alone.

[0017] The voltage model is a mathematical model based on the stator voltage equation of the asynchronous motor in the stationary coordinate system. Its basic principle is to calculate the flux linkage by integrating the back electromotive force of the motor. Since the back electromotive force can be calculated from the externally measured stator voltage and current and the known stator resistance, this voltage model is not coupled with the rotor speed, but serves as a reference model.

[0018] In its specific implementation, S2 first involves performing an inverse Parker transformation on the dq-axis stator current and voltage based on the estimated rotor magnetic field position to obtain the α-β-axis stator voltage and current. During this process, the rotor magnetic field position angle estimated in the previous control cycle is used to rotate the dq-axis stator voltage and current back to the two-phase stationary coordinate system, yielding the corresponding voltage and current components.

[0019] Next, the stator flux linkage is estimated using a compensated integrator based on the α-β axis stator voltage and α-β axis stator current to obtain the stator flux linkage. It should be understood that in the α-β stationary coordinate system, the time derivative of the stator flux linkage is equal to the back electromotive force (EMF). Theoretically, the stator flux linkage can be obtained by pure integration of the back EMF. However, pure integrators suffer from DC drift and initial value problems, and are prone to saturation in practical digital systems. Therefore, in the technical solution of this invention, a compensated integrator (usually a first-order low-pass filter) is used to replace the pure integrator. This effectively suppresses DC drift while ensuring integral characteristics, thereby stably and accurately calculating the stator flux linkage components on the α and β axes. In specific implementation, the α-β axis stator voltage and α-β axis stator current obtained in the previous step, along with the known stator resistance value, are substituted into the stator voltage equation, and the α-β axis components of the stator flux linkage are calculated using a compensated integrator (rather than an ideal integrator), thereby effectively avoiding the integral drift problem and ensuring the stability and accuracy of the calculation.

[0020] Furthermore, based on the stator flux linkage, α-β axis stator current, and motor electromagnetic parameters, the final algebraic derivation of the rotor flux linkage is performed to obtain the reference rotor flux linkage. That is, after obtaining a stable estimated value for the stator flux linkage, the reference rotor flux linkage is directly calculated based on the algebraic relationship between the stator and rotor flux linkages in the motor's steady-state model. Specifically, the final algebraic derivation of the rotor flux linkage is performed using the following formula: ; in, For stator inductance, For mutual induction, For rotor inductance, For the current period Shaft stator current, For the stator flux linkage in the current cycle, The reference rotor flux linkage for the current cycle, This indicates the current cycle. The formula combines the physical quantities on the stator side (stator flux linkage, stator current) with the inherent parameters of the motor, and finally calculates the reference rotor flux linkage under the voltage model through direct algebraic operations.

[0021] Specifically, in step S3, the rotor resistance is identified online using a hybrid identification method for the dq-axis stator current and dq-axis stator voltage to obtain an estimated rotor resistance and an estimated rotor time constant. It should be understood that the operating environment temperature of wind turbines varies drastically, which can cause fluctuations in the resistance value of the motor rotor windings of up to 40%-50%. Rotor resistance is a key parameter determining the dynamic response characteristics of the motor—a core component of the rotor time constant. In traditional sensorless control schemes, deviations in the rotor time constant directly disrupt the matching between the reference model and the adjustable model in the Model Reference Adaptive System (MRAS), leading to serious errors in the estimation of flux linkage and speed. Especially in the low-speed operating region where the back EMF is weak, the negative impact of this parameter mismatch is amplified dramatically, easily triggering system oscillations or even runaway control. Therefore, to achieve truly accurate and reliable sensorless control, the technical solution of this invention performs accurate, real-time online identification of the rotor resistance and dynamically updates the rotor time constant based on this identification.

[0022] In specific implementation of S3, firstly, the dq-axis stator current is input into the slow channel to obtain the predicted rotor resistance value. It should be understood that the thermal dynamics of a motor is a slow-responding but well-defined physical process, with a time constant typically on the order of minutes. Therefore, by establishing a slow channel based on the motor's thermodynamic characteristics, the general trend of resistance drift caused by temperature changes can be fundamentally predicted. The predicted rotor resistance value is stable and convergent in the long term, effectively suppressing the unbounded drift that may occur in the fast channel. This ensures that the entire hybrid identification system can output a long-term accurate resistance estimate within a reasonable physical range under any operating condition, significantly enhancing the robustness and reliability of the algorithm. It is worth mentioning that in the embodiments of this invention, the slow channel specifically refers to an algorithm path whose dynamic response characteristics are designed to match the actual thermal response time constant of the motor. Therefore, its output changes slowly and is mainly used to track motor temperature changes caused by load and environmental variations, thereby predicting the trend of resistance.

[0023] Specifically, inputting the dq-axis stator current into the slow channel to obtain the predicted rotor resistance value includes: First, determining the thermal load proxy index based on the dq-axis stator current. The thermal load proxy index is a value that can be easily calculated from online measured electrical quantities, used to indirectly characterize the main heat source inside the motor, namely the copper loss in the windings. Specifically, in the motor vector control system, the dq-axis stator current is calculated and controlled in real time. Since the total copper loss of the motor is mainly caused by the stator current, and the square of the total amplitude of the stator current is proportional to the copper loss, the sum of the squares of the d-axis stator current and the q-axis stator current can be taken as the thermal load proxy index. This index directly reflects the heat generation power of the motor under the current operating conditions and is a direct input for subsequent temperature estimation.

[0024] Secondly, based on the thermal load surrogate index and ambient temperature, the rotor temperature of the previous cycle is dynamically iteratively updated to obtain the rotor temperature of the current cycle. Specifically, this process is implemented in software as a simplified dynamic thermal model, such as an equivalent first-order thermal network model. This model treats the motor as an object with specific heat capacity and thermal resistance. In each calculation cycle, the model receives the thermal load surrogate index (representing heat input) determined in the previous step and the ambient temperature (representing heat dissipation boundary conditions) measured by sensors or preset, and combines it with the rotor temperature estimate of the previous cycle stored in memory to calculate the rotor temperature of the current cycle through a discretized dynamic equation. This iterative process simulates the dynamic change of the motor's internal temperature with load and environmental variations.

[0025] Next, a temperature-resistance linear mapping is performed on the rotor temperature of the current cycle to obtain the predicted rotor resistance value. This temperature-resistance linear mapping is based on a fundamental physical law: the resistance of a metallic conductor (in this case, the copper or aluminum material of the motor rotor windings) changes approximately linearly with its temperature. This established physical relationship is the theoretical basis for the slow channel's ability to deduce the resistance value from the estimated temperature.

[0026] Next, the dq-axis stator current and dq-axis stator voltage are input into the fast channel to obtain the resistance error signal. It should be understood that while the slow channel can provide long-term stable rotor resistance predictions, its inherent inertia based on the thermal model prevents it from quickly responding to instantaneous resistance fluctuations caused by rapid changes in motor operating conditions. For example, under conditions of frequent start-stop, sudden load changes, or rapid speed changes, the rotor resistance may change rapidly, and the slow channel's response speed is insufficient to track these changes in time. In this situation, the fast channel becomes particularly important. It utilizes the motor's real-time electrical quantities (dq-axis stator current and voltage) to construct a signal highly sensitive to resistance errors, aiming to provide immediate and precise correction capabilities. This ensures that the entire system maintains the accuracy of flux linkage estimation and the stability of speed adaptation even during rapid dynamic processes, avoiding control performance degradation or even instability caused by resistance model lag. It is worth mentioning that, in the embodiments of the present invention, the fast channel specifically refers to an algorithm path whose dynamic response characteristics are designed to react quickly. It is mainly used to capture instantaneous deviations in motor parameters (especially rotor resistance) and generate an error signal for correction, which makes up for the slow response of the slow channel.

[0027] Specifically, the dq-axis stator current and dq-axis stator voltage are input into a fast channel to obtain the resistance error signal. This includes: First, based on the dq-axis stator current and dq-axis stator voltage, the d-axis transient back EMF component and the q-axis transient back EMF component are calculated. Specifically, in the dq rotating coordinate system of the asynchronous motor, the stator voltage equation describes the relationship between the stator voltage, stator resistance voltage drop, the rate of change of stator flux linkage, and the induced EMF caused by rotor rotation. In practical implementation, the transient back EMF component can be understood as the portion remaining after subtracting the voltage components caused by the stator resistance voltage drop and the rate of change of stator inductance from the measured dq-axis stator voltage. In this process, the d-axis transient back EMF component can be obtained by subtracting the product of the stator resistance and the d-axis stator current from the d-axis stator voltage, and then subtracting the product of the stator inductance and the rate of change of the d-axis stator current with respect to time. The same calculation logic applies to the q-axis, that is, the q-axis transient back EMF component can be obtained by subtracting the product of the stator resistance and the q-axis stator current from the q-axis stator voltage, and then subtracting the product of the stator inductance and the rate of change of the q-axis stator current with respect to time. The obtained d-axis and q-axis transient back EMF components collectively reflect the dynamic induced electromotive force information inside the motor.

[0028] Secondly, resistance-sensitive cross-coupling is applied to the d-axis transient back EMF components, the q-axis transient back EMF components, and the dq-axis stator current to obtain the resistance error signal. The core idea of ​​this resistance-sensitive cross-coupling is to construct an error signal that is highly sensitive to changes in rotor resistance but relatively insensitive to changes in other parameters (such as rotational speed). Through this coupling, pure resistance error information can be effectively separated from complex electrical dynamics.

[0029] Furthermore, the predicted rotor resistance and the resistance error signal are fused and output to obtain the estimated rotor resistance and the estimated rotor time constant. Here, it can be understood that the slow channel provides a long-term, accurate, drift-free physical reference (predicted value), while the fast channel provides a fast-responding, locally corrected error signal. By intelligently fusing these two signals with different characteristics, the potential drift problem of the fast channel can be effectively suppressed, while significantly improving the dynamic response capability of the slow channel. This ensures that high-precision and high-stability rotor resistance estimates can be obtained under all operating conditions, especially in the complex environment of wide temperature variations and low-speed operation of wind power converters. This accurate resistance estimate is a key input for the subsequent normal operation of the speed adaptive module, directly affecting the accuracy of motor flux linkage and speed estimation, thereby guaranteeing the performance and reliability of the entire control system.

[0030] Specifically, fusing and outputting the predicted rotor resistance value and the resistance error signal to obtain the estimated rotor resistance and the estimated rotor time constant includes: First, performing a resistance estimation value fusion update based on dual-channel input on the predicted rotor resistance value and the resistance error signal to obtain the estimated rotor resistance. Specifically, in an embodiment of the present invention, the resistance estimation value fusion update based on dual-channel input on the predicted rotor resistance value and the resistance error signal is performed using the following formula: ; in, This is the resistance error signal for the current cycle. For the estimated rotor resistance of the previous cycle, To predict channel tracking gain, To correct the channel, the adjustment factor is increased. The sampling period is This is the predicted value of the rotor resistance for the current cycle. This formula estimates the rotor resistance for the current cycle. The calculation principle is to establish the current cycle's rotor resistance estimate based on the previous cycle's estimate, and then adjust it using two correction terms. The first correction term is the prediction channel tracking gain, which calculates the difference between the slow channel prediction and the previous cycle's estimate, multiplied by a prediction channel tracking gain. This correction term slowly pulls the estimated value towards a long-term stable prediction, serving a calibration and anti-drift function. The second correction term is the correction channel correction term, which directly utilizes the resistance error signal generated by the fast channel and multiplies it by a correction channel correction gain. This correction term rapidly and finely adjusts the estimated value based on real-time electrical signal deviations. By adding these two correction terms to the previous cycle's estimated rotor resistance, the estimated rotor resistance for the current cycle can be obtained.

[0031] Furthermore, based on the estimated rotor resistance and rotor inductance, the estimated rotor time constant is determined. That is, after obtaining the most accurate estimated rotor resistance for the current cycle through the aforementioned fusion update algorithm, the corresponding estimated rotor time constant is calculated using the known, relatively constant motor parameter—rotor inductance. Specifically, the estimated rotor time constant can be obtained by calculating the ratio of rotor inductance to estimated rotor resistance. The estimated rotor time constant is a crucial parameter characterizing the rotor flux linkage establishment and decay response time. In sensorless control, accurately acquiring and using the estimated rotor time constant is essential for flux linkage estimation based on the current model. Temperature drift in rotor resistance can cause changes in the time constant, thus affecting estimation accuracy and system stability.

[0032] Specifically, in step S4, based on the estimated rotor time constant, flux linkage estimation and speed adaptation based on a current model are performed on the dq-axis stator current to obtain the estimated rotor mechanical angular velocity and the estimated rotor magnetic field position. It should be understood that in traditional motor control systems, physical sensors such as encoders are typically required to directly measure the motor's speed and rotor position. However, in applications such as wind power generation, the use of physical sensors increases the system's cost, size, and complexity. Furthermore, the sensors themselves and their connecting cables are susceptible to harsh environments, becoming potential points of failure and reducing the overall system reliability. Therefore, in the technical solution of this invention, a sensorless technology is employed. By constructing a mathematical model of the motor, easily measurable electrical quantities (such as stator current) are used to indirectly estimate mechanical quantities (speed and position) within the motor that are difficult to measure directly. Specifically, through a model that is insensitive to changes in motor parameters (especially rotor resistance) and exhibits stability in the low-to-medium speed range, the rotor mechanical angular velocity and magnetic field position are accurately estimated, thereby achieving high-performance field-oriented control without relying on physical sensors.

[0033] In its specific implementation, S4 firstly derives and estimates the rotor flux linkage using the current equation of the motor in the dq synchronous rotating coordinate system. The current model, based on the rotor-side voltage equation or equivalent circuit, takes the dq-axis stator current as the primary input and calculates it using known motor parameters (e.g., stator inductance, mutual inductance, and rotor time constant). In this process, estimating the rotor time constant directly determines the response characteristics and steady-state accuracy of the rotor flux linkage estimator. The rotor time constant, the ratio of rotor inductance to rotor resistance, characterizes the rate at which rotor flux linkage builds up and decays. By combining the latest and most accurate estimation of the rotor time constant, the current model observer can continuously estimate the d-axis and q-axis components of the rotor flux linkage by utilizing the time-varying characteristics of the dq-axis stator current through filtering or integration.

[0034] Secondly, the speed adaptive mechanism is another core element of step S4. It executes in parallel with the flux linkage estimation process, aiming to correct the output of the flux linkage estimator and estimate the rotor's mechanical angular velocity online and dynamically. Due to the uncertainty of parameters in the motor model (especially rotor resistance) and the estimation drift problem that may occur in the current model under specific operating conditions, simple current model estimation is difficult to continuously provide accurate speed and position information in highly dynamic or parameter-changing environments. In the technical solution of this invention, speed adaptation is a dynamic adjustment mechanism used to estimate and correct the rotor's mechanical angular velocity online. It constructs an error function (e.g., formed by the deviation between the flux linkage estimate and the expected value, or by the electromagnetic torque error) and drives a closed-loop feedback system. This closed-loop feedback system continuously adjusts the estimated rotor electric angular velocity to minimize the error function, thereby enabling the flux linkage estimator to accurately synchronize with the actual motion state of the rotor. This mechanism enhances the robustness of sensorless control under dynamic response and parameter changes. Through these two closely cooperating sub-processes, accurate rotor mechanical angular velocity and rotor magnetic field position information can be continuously provided without a speed sensor, laying a solid foundation for high-precision field-oriented control.

[0035] Specifically, in S5, field-oriented control is performed on the estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command to obtain the PWM gate drive signal. It should be understood that due to the complex dynamic coupling characteristics of asynchronous motors, traditional voltage / frequency (V / f) control struggles to achieve fast response and high-precision torque control. Field-oriented control, through clever coordinate transformation, decouples the mutually coupled current equations of the AC motor in the stator reference frame into a DC-like motor model in the dq coordinate system, which rotates synchronously with the rotor flux linkage. In this decoupled control mode, the d-axis current component can be directly used to control the rotor flux linkage, while the q-axis current component is directly used to control the motor torque. Both can be adjusted independently, thereby achieving independent linear control of flux linkage and torque, enabling the AC motor to achieve high dynamic response and high-precision control performance.

[0036] In this context, PWM gate drive signals refer to a series of digital square wave signals generated using pulse width modulation technology. These signals are used to control the turn-on and turn-off times of power electronic switching devices (such as IGBTs or MOSFETs), thereby precisely adjusting the average value and waveform of the output voltage. In wind power converters, PWM signals drive the inverter to generate the required three-phase AC voltage to control the motor's speed and torque. Gate drive signals specifically refer to the voltage or current signals used to turn the gate (or base) of power electronic devices on and off.

[0037] In specific implementation, S5 first receives external speed and flux linkage commands, which represent the desired operating state of the motor. Next, it compares the speed command with the estimated rotor mechanical angular velocity obtained in the previous steps to obtain the speed error. This speed error is fed into a proportional-integral (PI) controller. The output of the PI controller is the torque command (typically corresponding to the q-axis current reference value) to drive the motor to generate the required torque. Its calculation principle is to gradually reduce the error by proportionally amplifying and integrating the speed error over time, making the actual speed approach the commanded speed, while simultaneously outputting the required q-axis current component to control the torque. Similarly, it compares the externally set flux linkage command with the amplitude of the reference rotor flux linkage to calculate the flux linkage error. This flux linkage error is also fed into a PI controller, which calculates the required d-axis current reference value for the excitation component based on the proportional and integral action of the flux linkage error. Its calculation principle is similar to that of the speed loop, aiming to adjust the d-axis current component to keep the actual rotor flux linkage amplitude consistent with the commanded flux linkage amplitude.

[0038] Subsequently, after obtaining the d-axis and q-axis current reference values, the system enters the internal current control loop. In this process, firstly, the calculated d-axis and q-axis current reference values ​​are compared with the d-axis and q-axis components of the stator current input from an external system (e.g., measured by a current sensor and obtained through Park transformation), thereby calculating the d-axis and q-axis current errors respectively. Next, the d-axis and q-axis current errors are fed into their respective independent PI controllers. Based on their respective current errors, the PI controllers generate d-axis and q-axis voltage reference values, using proportional-integral action to quickly and effectively eliminate current errors, enabling the actual d-axis and q-axis currents to rapidly track their reference values. Furthermore, to further improve the decoupling performance of the current control and ensure control accuracy, decoupling terms are typically added here. These decoupling terms compensate for the inherent electromagnetic coupling effects between the d and q axes, such as voltage cross-coupling caused by synchronous electric angular velocity and stator inductance, as well as the influence of back EMF. By adding these compensation terms, the behavior of d-axis voltage controlling d-axis current and q-axis voltage controlling q-axis current can be made more independent, reducing mutual interference.

[0039] Subsequently, after generating the voltage reference values ​​for the d-axis and q-axis, since the motor is actually driven by three-phase AC voltage, it is necessary to convert these voltage commands in the dq synchronous rotating coordinate system back to the three-phase stationary coordinate system in the physical world. This conversion process is achieved through the Park inverse transformation. The core of the Park inverse transformation is to use the estimated rotor magnetic field position obtained in the previous steps as the rotation angle. This angle accurately describes the orientation of the synchronous rotating coordinate system relative to the stationary coordinate system, enabling the transformation to accurately map the voltage components of the d-axis and q-axis to the corresponding three-phase voltage commands.

[0040] Furthermore, the three-phase voltage commands obtained through the Park inverse transformation are input to the PWM generation module. The main task of this module is to generate PWM gate drive signals based on these voltage commands to control the switching elements in the power electronic converter (usually an inverter). Commonly used PWM generation algorithms include Space Vector Pulse Width Modulation (SVPWM) or Three-Phase Sinusoidal Pulse Width Modulation (SPWM). SVPWM works by synthesizing a space voltage vector using the three-phase voltage commands, and then selecting appropriate inverter switching states and their durations to make the average effect of the inverter output voltage as close as possible to this command vector. Finally, the PWM module generates a set of high-frequency digital square wave signals, which are directly sent to the inverter's gate drive circuit as PWM gate drive signals. These PWM gate drive signals precisely control the on and off times of the switching elements such as IGBTs or MOSFETs inside the inverter, thereby inverting the DC bus voltage of the wind power converter into a three-phase AC voltage whose amplitude, frequency, and phase can be precisely controlled, thus driving the motor and achieving precise control of the motor speed and torque.

[0041] In summary, the sensorless field-oriented control method for wind power converters according to embodiments of the present invention has been clarified. It introduces a precise online hybrid identification mechanism for rotor resistance, estimates rotor resistance and time constant in real time, and deeply integrates them into the flux estimation and speed adaptation processes, thereby effectively decoupling the contradiction between parameter identification and speed estimation. In this way, the control accuracy and stability of the wind power converter are significantly improved when facing complex environmental temperature fluctuations and low-speed operating conditions, effectively avoiding flux estimation deviations and system runaway, making the sensorless control technology more robust and reliable.

[0042] This invention also provides a sensorless magnetic field orientation control system for wind power converters.

[0043] Figure 3 This is a block diagram of a sensorless field-oriented control system for a wind power converter according to an embodiment of the present invention. Figure 3As shown, the sensorless magnetic field orientation control system 300 for a wind power converter according to an embodiment of the present invention includes: a coordinate transformation module 310, used to perform coordinate transformation on the three-phase stator current and three-phase stator voltage corresponding to the wind power converter to obtain the dq-axis stator current and dq-axis stator voltage; a flux linkage estimation module 320, used to perform flux linkage estimation based on a voltage model on the dq-axis stator current and dq-axis stator voltage to obtain a reference rotor flux linkage; an online rotor resistance hybrid identification module 330, used to perform online rotor resistance hybrid identification on the dq-axis stator current and dq-axis stator voltage to obtain an estimated rotor resistance and an estimated rotor time constant; a speed adaptation module 340, used to perform flux linkage estimation based on a current model and speed adaptation on the dq-axis stator current based on the estimated rotor time constant to obtain an estimated rotor mechanical angular velocity and an estimated rotor magnetic field position; and a magnetic field orientation control module 350, used to perform magnetic field orientation control on the estimated rotor mechanical angular velocity, the estimated rotor magnetic field position, the reference rotor flux linkage, the speed command, and the flux linkage command to obtain a PWM gate drive signal.

[0044] The specific implementation method of the sensorless magnetic field orientation control system for wind power converters provided in this embodiment of the invention can be found in the description of the sensorless magnetic field orientation control method for wind power converters provided in this embodiment of the invention, and will not be repeated here.

[0045] The sensorless field-oriented control system 300 for wind power converters according to embodiments of the present invention can be implemented in various wireless terminals, such as servers with sensorless field-oriented control algorithms for wind power converters. In one possible implementation, the sensorless field-oriented control system 300 for wind power converters according to embodiments of the present invention can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the sensorless field-oriented control system 300 for wind power converters can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the sensorless field-oriented control system 300 for wind power converters can also be one of many hardware modules of the wireless terminal.

[0046] Alternatively, in another example, the sensorless magnetic field orientation control system 300 for the wind power converter and the wireless terminal can also be separate devices, and the sensorless magnetic field orientation control system 300 for the wind power converter can be connected to the wireless terminal via wired and / or wireless networks, and transmit interactive information in accordance with an agreed data format.

[0047] Those skilled in the art will understand that the above embodiments are specific implementations of the present invention, and in practical applications, various changes can be made in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A sensorless field-oriented control method for wind power converters, characterized in that, include: The three-phase stator current and three-phase stator voltage of the wind power converter are transformed by coordinate to obtain the dq-axis stator current and dq-axis stator voltage; The flux linkage is estimated based on the voltage model for the dq-axis stator current and dq-axis stator voltage to obtain the reference rotor flux linkage; Online hybrid identification of rotor resistance is performed on the dq-axis stator current and dq-axis stator voltage to obtain estimated rotor resistance and estimated rotor time constant; Based on the estimated rotor time constant, flux linkage estimation and speed adaptation based on the current model are performed on the dq-axis stator current to obtain the estimated rotor mechanical angular velocity and the estimated rotor magnetic field position. The estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command are used to perform field-oriented control to obtain the PWM gate drive signal.

2. The sensorless magnetic field orientation control method for wind power converters according to claim 1, characterized in that, The reference rotor flux linkage is obtained by estimating the stator current and stator voltage along the dq axis based on a voltage model, including: Based on the estimated rotor magnetic field position, an inverse Parker transformation is performed on the dq-axis stator current and dq-axis stator voltage to obtain the α-β-axis stator voltage and α-β-axis stator current; The stator flux linkage is estimated by using a compensated integrator based on the α-β axis stator voltage and α-β axis stator current. The final algebraic derivation of the rotor flux linkage is performed based on the stator flux linkage, α-β axis stator current, and motor electromagnetic parameters to obtain the reference rotor flux linkage.

3. The sensorless magnetic field orientation control method for wind power converters according to claim 2, characterized in that, The final algebraic derivation of the rotor flux linkage based on the stator flux linkage, α-β axis stator currents, and motor electromagnetic parameters is used to obtain the reference rotor flux linkage, including: The final algebraic derivation of the rotor flux linkage is performed using the following formula: ； in, For stator inductance, For mutual induction, For rotor inductance, For the current period Shaft stator current, For the stator flux linkage in the current cycle, The reference rotor flux linkage for the current cycle, Indicates the current period.

4. The sensorless magnetic field orientation control method for wind power converters according to claim 1, characterized in that, Online hybrid identification of rotor resistance is performed on the dq-axis stator current and dq-axis stator voltage to obtain estimated rotor resistance and estimated rotor time constant, including: Input the dq-axis stator current into the slow channel to obtain the predicted value of the rotor resistance; Input the dq-axis stator current and dq-axis stator voltage into the fast channel to obtain the resistance error signal; The predicted rotor resistance and the resistance error signal are fused and output to obtain the estimated rotor resistance and the estimated rotor time constant.

5. The sensorless magnetic field orientation control method for wind power converters according to claim 4, characterized in that, Inputting the dq-axis stator current into the slow channel yields the predicted rotor resistance value, including: Determine the thermal load proxy index based on the dq axis stator current; Based on the thermal load proxy index and ambient temperature, the rotor temperature of the previous cycle is dynamically iterated and updated to obtain the rotor temperature of the current cycle. A temperature-resistance linear mapping is performed on the rotor temperature of the current cycle to obtain the predicted rotor resistance value.

6. The sensorless magnetic field orientation control method for wind power converters according to claim 4, characterized in that, The dq-axis stator current and dq-axis stator voltage are input into the fast channel to obtain the resistance error signal, including: Based on the dq-axis stator current and dq-axis stator voltage, calculate the d-axis transient back EMF component and the q-axis transient back EMF component; The d-axis transient back EMF component, the q-axis transient back EMF component, and the dq-axis stator current are subjected to resistance-sensitive cross-coupling to obtain the resistance error signal.

7. The sensorless magnetic field orientation control method for wind power converters according to claim 4, characterized in that, The predicted rotor resistance and the resistance error signal are fused and output to obtain the estimated rotor resistance and the estimated rotor time constant, including: The rotor resistance prediction value and the resistance error signal are fused and updated based on dual-channel input to obtain the estimated rotor resistance; The estimated rotor time constant is determined based on the estimated rotor resistance and rotor inductance.

8. The sensorless magnetic field orientation control method for wind power converters according to claim 7, characterized in that, The rotor resistance prediction value and the resistance error signal are fused and updated based on dual-channel input to obtain the estimated rotor resistance, including: The rotor resistance prediction and resistance error signal are fused and updated based on dual-channel input using the following formula: ； in, This is the resistance error signal for the current cycle. For the estimated rotor resistance of the previous cycle, To predict channel tracking gain, To correct the channel, the adjustment factor is increased. The sampling period is This is the predicted value of the rotor resistance for the current cycle. This is the estimated rotor resistance for the current cycle.

9. A sensorless magnetic field orientation control system for a wind power converter, characterized in that, include: The coordinate transformation module is used to perform coordinate transformation on the three-phase stator current and three-phase stator voltage corresponding to the wind power converter to obtain the dq-axis stator current and dq-axis stator voltage; The flux linkage estimation module is used to perform voltage model-based flux linkage estimation on the dq-axis stator current and dq-axis stator voltage to obtain the reference rotor flux linkage. The online hybrid identification module for rotor resistance is used to perform online hybrid identification of rotor resistance based on the dq-axis stator current and dq-axis stator voltage to obtain an estimated rotor resistance and an estimated rotor time constant. The speed adaptive module is used to perform flux linkage estimation and speed adaptation based on the current model of the dq-axis stator current based on the estimated rotor time constant, so as to obtain the estimated rotor mechanical angular velocity and the estimated rotor magnetic field position. The field orientation control module is used to perform field orientation control on the estimated rotor mechanical angular velocity, estimated rotor magnetic field position, reference rotor flux linkage, speed command, and flux linkage command to obtain the PWM gate drive signal.

Images