Motor state monitoring method based on similar stochastic resonance reaching law sliding mode observer
By designing a sliding mode observer based on the stochastic resonance approach law, the problem of jitter in the event of failure of the double-feed induction motor is solved, fast convergence and high tracking accuracy are achieved, and robustness and applicability are improved.
Patent Information
- Application Number
- CN202411459332.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-18
- Publication Date
- 2025-06-17
AI Technical Summary
Dual-feed induction motors are prone to jitter problems due to failures in wind power generation systems, which affects the robustness and observation accuracy of the observer. The existing approach law design is difficult to reduce jitter and improve convergence speed at the same time.
A sliding mode observer based on the approaching law of random resonance is designed. By improving the power approaching law and combining the signal noise reduction capability of random resonance, a sliding mode observer suitable for double-feed induction motors is constructed.
It realizes rapid convergence and high tracking accuracy of the sliding mode observer when the system is normal and faulty, reduces vibration and improves robustness and applicability.
Smart Images
Figure CN120161335A_ABST
Abstract
Description
[0001] The present invention belongs to the field of motor condition monitoring, and specifically provides a method for monitoring the condition of a doubly-fed induction motor based on a sliding mode observer with a quasi-random resonance reaching law. Background Art
[0002] At present, wind power generation is developing very rapidly and has a wide range of applications. As a type of motor widely used in wind power generation systems, doubly-fed induction motors are susceptible to various motor faults due to the harsh external environment of wind power generation. However, there are no obvious characteristics in the early stage of motor faults, which may lead to significant economic losses and even casualties after the faults deteriorate. Therefore, how to monitor the condition of the motor and ensure accurate observation when the motor fails has become the research focus in this field.
[0003] Benefiting from the superiority of the sliding mode variable structure control theory applied to nonlinear systems, sliding mode observers are widely used in the fields of motor condition monitoring and motor fault diagnosis. However, the sliding mode observer designed based on the reaching law method is prone to chattering problems, which will affect the robustness and observation accuracy of the observer. The current mainstream method is to design a new reaching law to reduce chattering while ensuring the convergence speed of the observer. However, it is difficult to satisfy the above two results simultaneously. Although some reaching laws in existing research can take both reducing chattering and improving the convergence speed into account, the application scenarios are not complex nonlinear systems such as motors. Therefore, it has certain research significance to design a sliding mode observer with high applicability and good tracking accuracy for doubly-fed induction motors by using the reaching law method to achieve the condition monitoring of the motor. Summary of the Invention
[0004] Based on the problems in the above background, the present invention designs a method for monitoring the condition of a motor based on a sliding mode observer with a quasi-random resonance reaching law. The power reaching law is improved by using the signal denoising ability of stochastic resonance. The improved reaching law can not only reduce chattering but also improve the convergence speed, making the designed sliding mode observer have better applicability and higher tracking accuracy.
[0005] The specific solution of the present invention is as follows: Step 1: Improve the power reaching rate by using the signal denoising ability of stochastic resonance, design a quasi-random resonance reaching law, and construct a sliding mode observer based on the state equation of the doubly-fed induction motor in combination with the improved reaching law; Step 2: Use the Lyapunov stability theory to analyze the stability of the designed reaching law, define the value range of the gain parameters in the reaching law, and analyze the anti-disturbance performance of the designed sliding mode observer; Step 3: Establish a simulation model of the doubly-fed induction motor under variable wind speed faults, input interference voltage faults, voltage sag faults, stator inter-turn short circuit faults, and rotor current sensor faults. Compare the designed observer with the sliding mode observers designed by traditional exponential reaching laws and power reaching laws. Use the residual values between the true value of the d-axis component of the rotor current output by the motor and the estimated values output by each sliding mode observer to verify the versatility and state tracking effect of the designed observer.
[0006] Advantages of the present invention:
[0007] (1) The state monitoring method of the doubly-fed induction motor proposed by the present invention mainly relies on a sliding mode observer based on a class of stochastic resonance reaching law. Specifically, on the basis of the power reaching law, after being improved by stochastic resonance, it can not only improve the overall convergence speed but also reduce the degree of change in the convergence rate when approaching the sliding surface, thereby reducing chattering. Furthermore, when the system is normal and faulty, the designed observer has a fast convergence speed and excellent tracking accuracy.
[0008] (2) The designed new sliding mode observer has good applicability to several common electrical faults such as variable wind speed faults, voltage sag faults, input interference voltage faults, stator winding inter-turn short circuit faults, and rotor current sensor faults. It provides an effective method for the research on the state monitoring of wind power generation systems based on doubly-fed induction motors. Description of the drawings
[0009] Figure 1 It is a schematic diagram of the state monitoring process of a doubly-fed induction motor based on a sliding mode observer with a class of stochastic resonance reaching law.
[0010] Figure 2 It is a comparison diagram of the stochastic resonance term and the power reaching law term.
[0011] Figure 3 It is a comparison diagram of the class of stochastic resonance reaching law with the stochastic resonance term and the power reaching law term.
[0012] Figure 4 It is a flow chart of the comparative experiment of the class of stochastic resonance sliding mode observer with two traditional sliding mode observers.
[0013] Figure 5 It is a schematic diagram of the tracking effect of the true value and the estimated value of the d-axis component of the rotor current under variable wind speed faults.
[0014] Figure 6 It is a schematic diagram of the residual between the true value and the estimated value of the d-axis component of the rotor current under variable wind speed faults.
[0015] Figure 7 It is a schematic diagram of the tracking effect of the true value and the estimated value of the d-axis component of the rotor current under input interference voltage faults.
[0016] Figure 8 It is a schematic diagram of the residual between the true value and the estimated value of the d-axis component of the rotor current under the input interference voltage fault. Specific implementation scheme
[0017] In order to describe the present invention more clearly, the present invention will be described in detail below with reference to the accompanying drawings.
[0018] A motor state monitoring method based on a quasi-random resonance reaching law sliding mode observer proposed by the present invention has a process as Figure 1 shown, and specifically includes the following steps: Step 1: Utilize the signal noise reduction ability of stochastic resonance to improve the power reaching law, design a quasi-random resonance reaching law, and based on the state equation of a doubly-fed induction motor, combine the improved reaching law to construct a sliding mode observer.
[0019] The expression of the quasi-random resonance reaching law in Step 1 is:
[0020] In the above formula, are all greater than 0, and . represents the sliding mode surface, and , where in the formula is the sliding mode surface gain parameter, and .
[0021] The stochastic resonance potential function of the classical single-well is , Then its derivative can be obtained as: , . Thanks to the ability of stochastic resonance to reduce signal noise, the present invention organically combines stochastic resonance with the traditional power reaching law to achieve the effects of improving the convergence speed and suppressing chattering.
[0022] The stochastic resonance term designed based on the derivative of the stochastic resonance potential function of the classical single-well has a specific expression of: , . In order to reduce chattering, the sign function in the traditional power reaching law is replaced with the hyperbolic tangent function, and the power reaching law term is designed as: , where . When , , and The images of Figure 2 are as shown. It can be clearly seen that The overall slope is greater than , indicating that when the numerical change is the same, Compared with has a greater numerical change. Then, for the observer, when it is far from the sliding mode surface , the faster the change, the faster the estimated value output by the observer will converge to the true value; when approaching the sliding mode surface , the slope change degree of is more severe than that of
[0023] which means that after the estimated value output by the observer converges to the true value, the change fluctuation around the true value is more obvious, and the resulting chattering is more severe. Therefore, based on the above discussion, the described quasi-random resonance reaching law is obtained. Figure 3 The comparison graph of the improved reaching law and the first two items is as shown in It can be clearly seen that after adding the random resonance term, the curve of the power reaching law term shrinks inward as a whole, resulting in a larger slope. Furthermore, it makes: When the state variable and are far from the sliding mode surface, and in the designed reaching law are the dominant terms, and as the numerical difference between becomes larger, the overall slope of becomes larger, thereby improving the convergence speed; When the state variable and are close to the sliding mode surface, and and in the designed reaching law are the dominant terms, which reduces the slope change degree of Figure 3 and can be clearly seen from the reference that while reducing the convergence speed, the chattering is alleviated. Moreover, as the numerical difference between
[0024] The expression of the sliding mode observer in the first step is:
[0025] According to the rotor current state equation of the doubly-fed induction motor, the described quasi-random resonance reaching law is introduced to construct a sliding mode observer with the motor rotor current as the state variable. The specific method is as follows: The rotor current state equation of the doubly-fed induction motor is:
[0026] where , , , .
[0027] In the above formula, is the input voltage variable, and the expression is: , where represent the d-q axis components of the stator and rotor voltages respectively; , represents the d-q axis components of the motor rotor current; , represents the derivative of the d-q axis components of the motor rotor current; , represents the estimated value of the d-q axis components of the motor rotor current by the sliding mode observer; , are the resistances of the stator winding and the rotor winding respectively; , are the inductances of the stator winding and the rotor winding respectively, is the mutual inductance between the stator winding and the rotor winding; is the stator flux linkage, are the stator speed and the rotor speed respectively; is the relative speed between the stator and the rotor, and ; ; The present invention adopts the stator voltage orientation control method. When the stator resistance is ignored, the equations satisfied by the stator voltage and the stator flux linkage are:
[0028] In the above formula is the amplitude of the stator flux linkage; Based on the above formula, the relationship between the stator current and the rotor current is established as follows:
[0029] In the above formula represents the stator magnetic flux. Substituting the above formula into the rotor current state equation, we can get:
[0030] In the above formula, , , .
[0031] Construct a sliding mode observer according to the rotor current state equation as follows:
[0032] In the above formula represents the estimated value of the rotor current, is the control law of the sliding mode observer, and the expansion is expressed as , , .
[0033] Set the sliding mode switching surface as follows:
[0034] Wherein, , , and are the sliding mode surface parameters.
[0035] Combine the designed quasi-stochastic resonance reaching law with the sliding mode variable structure control theory to design the sliding mode control law as follows:
[0036] Substitute the sliding mode control law into the rotor current sliding mode observer model to obtain the sliding mode observer based on the quasi-stochastic resonance reaching law as follows:
[0037] Step 2: Use the Lyapunov stability theory to analyze the stability of the designed reaching law, define the value range of the gain parameters in the reaching law, and analyze the disturbance rejection performance of the designed sliding mode observer.
[0038] Analyze the stability of the quasi-stochastic resonance reaching law as follows: For the quasi-stochastic resonance reaching law, it is necessary to use the Lyapunov stability theory to test its stability. Set the Lyapunov function as: , where . Then, the derivative of the function is as follows:
[0039] To make the above formula satisfy the Lyapunov stability theory, that is Then there is: . Therefore, for the quasi-stochastic resonance reaching law proposed in the present invention, when are all greater than 0 and , it can ensure that the Lyapunov stability theory is satisfied, and further ensure the robustness of the designed sliding mode observer. And, according to the analysis of the designed reaching law, as increases, the designed reaching law will increase the chattering while increasing the convergence speed. Therefore, set the value range of the gain parameter as [8000, 10000], the value range of
[0040] Analyze the stability of the observer as follows: In view of the unstable interference that may occur in the doubly-fed induction motor, it is necessary to analyze the robustness of the designed observer to the rotor current when subjected to external disturbances and to achieve rotor current tracking under different states.
[0041] Let the external disturbance be ,and , combined with the state equation of the doubly fed induction motor, we can get:
[0042] In the above formula, is the uncertainty interference part of the doubly fed wind turbine model, and the uncertainty interference part satisfies the following formula:
[0043] Combining the motor state equation with external disturbance and the mathematical model of the designed observer, we can get:
[0044] Combining the set sliding surface equation with the above formula, we can get:
[0045] According to Lyapunov stability theory, when uncertainty disturbance occurs in the doubly fed wind turbine , if the sliding mode observer wants to accurately track the system state, the conditions that must be met are . Then we have:
[0046] The above formula must satisfy Lyapunov stability theory, then:
[0047] In the above formula, The size of The impact on and Two situations, There are different values. So the process of reaching system stability can be divided into two stages for discussion.
[0048] (1) Order hour, The initial value is ,for arrive The stages are:
[0049] so The value range of is: .
[0050] (2) For to The stages are as follows:
[0051] Therefore The value range of is:
[0052] Based on the above discussion The defined range of is satisfies When the external disturbance
[0053] Step 3: Establish a simulation model of the doubly-fed induction motor under variable wind speed fault, input interference voltage fault, voltage sag fault, stator inter-turn short circuit fault, and rotor current sensor fault. Compare the designed observer with the sliding mode observers designed by the traditional exponential reaching law and power reaching law. Use the residual value between the true value of the d-axis component of the rotor current output by the motor and the estimated value output by each sliding mode observer to verify the versatility and state tracking effect of the designed observer.
[0054] For the five faults of variable wind speed fault, input interference voltage fault, voltage sag fault, stator inter-turn short circuit fault, and rotor current sensor fault, the settings are as follows:
[0055] (1) Variable wind speed fault As a common type of machine in wind power generation systems, the variable wind speed fault is one of the more common faults. Set the occurrence times of sudden wind speed changes to 0.4 s and 0.8 s. Specifically The settings of the fault value changes are as follows:
[0056] (2) Input interference voltage fault Since the wind power generation system based on the doubly-fed induction motor is directly connected to the power grid, the interference voltage from the power grid is likely to affect the operating state of the motor. Set the occurrence times of the fault to 1 s and 1.5 s. Specifically The fault settings are as follows:
[0057] (3) Voltage sag fault Since both the rotor side and the stator side of the doubly-fed induction motor are directly connected to the power grid, when a A, B, C three-phase grounding fault occurs in the power grid, it will cause a voltage sag fault in the power grid, which will further harm the normal operation of the motor. Set the fault to occur between 0.4 s and 0.8 s. Then the space vector expression of this fault is as follows:
[0058] Among them represents the amplitude of the grid voltage, represents the electrical angular velocity of the grid voltage, represents the percentage of the fault degree.
[0059] (4) Stator inter-turn short circuit fault As a frequently occurring fault type in motors, the stator inter-turn short circuit fault often brings great harm. In this paper, a sudden fault is used to simulate the stator inter-turn short circuit fault, and the fault occurrence time is set to 0.5 s. The specific fault is set as follows:
[0060] (5) Rotor current sensor fault Rotor current sensor faults are very common in wind power generation systems based on doubly-fed induction motors, usually occurring at the output end of the system. If not handled properly, they will affect the overall system. The occurrence times of this fault are set to 0.4 s, 0.6 s, and 0.8 s. The specific fault design is as follows:
[0061] The mathematical expressions of the sliding mode observers designed by the traditional exponential reaching law and power reaching law are as follows:
[0062] The specific experimental process is as Figure 4 shown. In this experiment, the superiority of the quasi-stochastic resonance sliding mode observer designed in the present invention will be verified by comparing the residual values between the true value of the d-axis component of the output rotor current of the motor under different faults and the estimated values output by each sliding mode observer. In each fault simulation experiment, the quasi-stochastic resonance reaching law sliding mode observer (QSRRL-SMO) designed in the present invention, the traditional exponential reaching law sliding mode observer (ERL-SMO), and the power reaching law sliding mode observer (PRRL-SMO).
[0063] The present invention only shows the experimental result diagrams under variable wind speed faults and input interference voltage faults. The remaining three fault results are similar and will not be shown. The specific experimental results are as Figures 5 to 8 shown, where Figure 5 and Figure 7 respectively represent the tracking situations of the true value of the d-axis component of the output rotor current of the motor and the estimated values output by each sliding mode observer under variable wind speed faults and input interference voltage faults. It can be seen from the images that the quasi-stochastic resonance sliding mode observer proposed in the present invention has better tracking effects than the other two traditional sliding mode observers, whether in the system stable stage or the fault stage.Figure 6 and Figure 8 respectively represent the residual situations of the true value and the estimated value output by each sliding mode observer of the d-axis component of the motor output rotor current during variable wind speed faults and input interference voltage faults. According to the image and combined with the numerical values in Table 1 and Table 2, it can be concluded that the quasi-random resonance sliding mode observer proposed in the present invention has the smallest value during the stable stage of the system; during the fault stage, compared with the other two traditional sliding mode observers, the new observer has smaller fluctuations and a faster convergence speed.
[0064]
[0065]
Claims
1. A motor state monitoring method based on a stochastic resonance reaching law sliding mode observer is mainly characterized by comprising the following steps: Step 1: Use the signal noise reduction capability of stochastic resonance to improve the power-order reaching rate, design a stochastic resonance-like reaching law, and construct a sliding mode observer based on the state equation of the doubly-fed induction motor combined with the improved reaching law; Step 2: Using Lyapunov stability theory, the designed reaching law is analyzed for stability, the value range of the gain parameter in the reaching law is defined, and the anti-disturbance analysis of the designed sliding mode observer is performed; Step 3: Establish a simulation model of the doubly-fed induction motor under variable wind speed fault, input interference voltage fault, voltage drop fault, stator turn-to-turn short circuit fault and rotor current sensor fault, compare the designed observer with the sliding mode observer designed by the traditional exponential reaching law and power reaching law, and use the residual value of the true value of the d-axis component of the motor output rotor current and the estimated value output by each sliding mode observer to verify the versatility and state tracking effect of the designed observer.
2. The motor state monitoring method based on the stochastic resonance reaching law sliding mode observer according to claim 1 is mainly characterized in that the expression of the stochastic resonance reaching law in the step 1 is: ， In the above formula, are greater than 0, and ; represents the sliding surface, and , where is the sliding surface gain parameter, and ; The stochastic resonance term designed based on the derivative of the stochastic resonance potential function of the classical single potential well is expressed as follows: , ; In order to reduce chattering, the sign function in the traditional power reaching law is replaced by the hyperbolic tangent function, and the power reaching law term is designed as: ,in ;when , When The overall slope is greater than , indicating that When the value changes are the same, Compared to has a larger numerical change; then for the observer, when it is far away from the sliding surface hour, The faster the change, the faster the estimated value of the observer output will converge to the true value; when it is close to the sliding surface hour, The slope change ratio It is more intense, indicating that after the estimated value of the observer output converges to the true value, the fluctuation around the true value is more obvious, and the resulting chattering is more intense; therefore, based on the above discussion, the stochastic resonance convergence law is obtained; (1) When the state variable When moving away from the sliding surface, the designed reaching law and As the leading item, and with and The greater the difference in the values of The overall slope becomes larger, thereby increasing the convergence speed; (2) When the state variable When approaching the sliding surface, the designed reaching law and As the leading item, The slope of the curve is reduced, which reduces the chattering while reducing the convergence speed. and The greater the difference in the values, the greater the vibration will be.
3. The method for state monitoring of a doubly-fed induction motor based on a quasi-stochastic resonance reaching law sliding mode observer according to claim 1, wherein the expression of the sliding mode observer in step 1 is: ， In the above formula, is the input voltage variable, and the expression is: ,in Represent the dq axis components of the stator and rotor voltages respectively; , Represents the dq-axis component of the motor rotor current; , Represents the derivative of the dq-axis component of the motor rotor current; , The sliding mode observer estimates representing the dq-axis components of the motor rotor current; , , , , is the sliding surface parameter, and .