Single-phase pulse width modulation rectifier sensing system fault feature extraction and active blocking control method

Abstract

The invention discloses a fault feature extraction and active blocking control method for a single-phase pulse width modulation rectifier sensing system, and the method comprises the steps: S1, constructing an unknown input sliding-mode observer, and carrying out the online observation of a state variable; s2, designing an unknown input variable estimation scheme, and performing parameter setting on the unknown input sliding mode observer; and S3, fault features of the single-phase pulse width modulation rectifier are extracted, and an active blocking control system is established. According to the control method provided by the invention, the defects of high complexity, poor portability, dependence on hardware redundancy and the like in the existing method are overcome, and the emergency survival capability of the single-phase pulse width modulation rectifier under various sensing system fault conditions is improved.

Description

Technical Field

[0001] The invention relates to a power electronic system control technology, in particular to a fault feature extraction and active blocking control method for a single-phase pulse width modulation rectifier sensor system. Background Art

[0002] The single-phase pulse width modulation rectifier circuit has a simple circuit topology and good controllability. It has been widely used in the fields of electrified rail transit, new energy vehicles, etc. However, in the actual working environment, due to the influence of factors such as high temperature, high humidity, chemical corrosion, and static electricity, the sensor detection part of the single-phase pulse width modulation rectifier often fails. For the control of the single-phase pulse width modulation rectifier, this sensor abnormality or loss will make it impossible to achieve effective closed-loop control of the grid-side current and the DC-side voltage, thereby endangering the overall operation stability and safety of the industrial control system it carries. In recent years, domestic and foreign scholars have conducted in-depth research on the fault diagnosis and fault-tolerant control of power converter sensor systems. The mainstream methods are divided into data-driven methods and analytical model-based methods.

[0003] The fault diagnosis strategies based on data-driven algorithms such as artificial neural networks proposed in literature (B.Gou, Y.Xu, Y.Xia, et al. An Intelligent Time-Adaptive Data-Driven Method for Sensor Fault Diagnosis in Induction Motor Drive System [J]. IEEE Transactions on Industrial Electronics, 2019, 66(12): 9817-9827) can achieve accurate diagnosis of various faults in power electronic systems. However, such methods are highly complex and can only be used for fault detection, and cannot provide reliable active blocking control solutions. The solution based on the Romberg observer proposed in the literature (J.Xia, Z.Li, Y.Guo, et al. Adaptive-Observer-Based Sensor Fault Resilient Control for Single-Phase Grid-Connected Converters in High-Speed ​​Railway Traction Systems [C] / / Proceedings of 2021 IEEE Energy Conversion Congress and Exposition. IEEE, 2021: 5190-5195) is only applicable to the bipolar modulation condition of the single-phase pulse width modulation rectifier, and the control reconstruction requires shutdown. In addition, in order to realize the observation of the grid-side current and DC-side voltage of the single-phase pulse width modulation rectifier, this type of method needs to introduce a redundant load current sensor, which increases the system cost.

[0004] Through the analysis of existing research results, it can be seen that the design process of the data-driven method is relatively cumbersome, and two different algorithms need to be designed to realize fault feature extraction and active blocking control respectively. The accuracy of industrial digital microprocessor signal acquisition and the speed of signal processing are high, which is not easy to be applied to single-phase pulse width modulation rectifiers; and the existing analytical model-based method requires the introduction of additional sensors, which increases the system cost. Therefore, how to achieve simple and integrated fault feature extraction and active blocking control for different sensor system fault types without introducing additional redundant sensors has become the key to improving the operating reliability of single-phase pulse width modulation rectifiers. Summary of the invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a method for extracting fault features and actively blocking control of a single-phase pulse width modulation rectifier sensor system, so as to overcome the shortcomings of the existing methods such as high complexity, poor portability, and reliance on hardware redundancy, and to improve the emergency survivability of the single-phase pulse width modulation rectifier under various sensor system failure conditions.

[0006] Technical solution: The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method described in the present invention comprises the following steps:

[0007] S1. Constructing an unknown input sliding mode observer to observe state variables online; including: establishing a state equation of a single-phase pulse width modulation rectifier, using a symbolic function to design a parameter matrix, and constructing an unknown input sliding mode observer to achieve online observation of state variables;

[0008] S2. Design an estimation scheme for unknown input variables and perform parameter tuning for the unknown input sliding mode observer; including: designing an online estimation method for unknown input variables of a single-phase pulse width modulation rectifier, constructing a Lyapunov function based on the quadratic form of the state variable observation error, and obtaining the lower limit of the coefficient of the symbolic function term of the unknown input sliding mode observer by scaling its time derivative;

[0009] S3. Extract the fault characteristics of the single-phase pulse width modulation rectifier and establish an active blocking control system; including: selecting the normalized value of the state variable observation error as the residual, extracting the fault characteristics of the single-phase pulse width modulation rectifier, realizing online detection and positioning of the fault sensing system, and using the observation value of the unknown input sliding mode observer to replace the measurement value of the fault sensing system to realize active blocking control until the fault is eliminated.

[0010] Furthermore, in step S1, a state equation of a single-phase pulse width modulation rectifier is established, specifically:

[0011] The dynamic equation of the single-phase pulse width modulation rectifier is established as:

[0012]

[0013] Among them, R is the grid side resistance, L is the grid side inductance, u s is the grid-side voltage, i s is the grid-side current, i L is the load current, u dc is the DC side voltage, S A With S B is the equivalent switching function of the two bridge arms, C is the equivalent capacitance on the DC side;

[0014] The established state equation is:

[0015]

[0016] in, is the time derivative of the state variable, x is the state variable, y is the output variable, u1 is the known input variable, u2 is the unknown input variable, A is the parameter matrix, B1 is the parameter matrix corresponding to the known input variable, B2 is the parameter matrix corresponding to the unknown input variable, and C is the output parameter matrix.

[0017] Furthermore, in step S1, a parameter matrix is ​​designed by using a symbolic function to construct an unknown input sliding mode observer to achieve online observation of state variables, specifically:

[0018] Define the state variable observation error vector as in is the observed value of the state variable, and x is the state variable;

[0019] Define the nonlinear feedback matrix E(e) to satisfy:

[0020]

[0021] Among them, η is the coefficient to be designed, η≥ρ>0, ρ is a positive number, F is the coefficient matrix, and C is the output parameter matrix;

[0022] The nonlinear feedback matrix E(e) is expressed in the form of a symbolic function and transformed into a function of the output variable and its observed value:

[0023]

[0024] Among them, y is the output variable, is the observed value of the output variable;

[0025] Therefore, the unknown input sliding mode observer is designed as:

[0026]

[0027] in, is the time derivative of the observed value of the state variable, u1 is the known input variable, A is the parameter matrix, B1 is the parameter matrix corresponding to the known input variable, B2 is the parameter matrix corresponding to the unknown input variable, and G is the gain matrix of the output error feedback term;

[0028] The state variable x is observed online based on the designed unknown input sliding mode observer.

[0029] Furthermore, step S2 is specifically as follows:

[0030] The estimation method for unknown input variables in the design of a single-phase pulse width modulation rectifier is:

[0031]

[0032] in, is the observed value of the unknown input variable, η is the coefficient to be designed, η≥ρ>0, ρ is a positive number, F is the coefficient matrix, y is the output variable, is the observed value of the output variable; in order to reconstruct the unknown input variable, a low-pass filter is used to filter it;

[0033] The dynamic equation of the state variable observation error vector e is expressed as:

[0034]

[0035] in, is the time derivative of the state variable observation error, is the time derivative of the observed value of the state variable, is the time derivative of the state variable, A is the parameter matrix, B2 is the parameter matrix corresponding to the unknown input variable, C is the output parameter matrix, H is the gain matrix, E(e) is the nonlinear feedback matrix, and u2 is the unknown input variable;

[0036] When the state variables are observable, there exists a gain matrix H and symmetric positive definite matrices P and Q such that (A-HC) T P+P(A-HC)=-Q holds; where T represents matrix transpose;

[0037] The observer parameters are adjusted using the method based on the Lyapunov function. The Lyapunov function V is:

[0038] V=e T P

[0039] Since the single-phase pulse width modulation rectifier has only one unknown input variable, namely the load current, the vector form of u2 is simplified to a variable u2, and the variable σ is defined to satisfy The time derivative of the Lyapunov function is obtained and simplified to:

[0040]

[0041] Among them, e T The Qe term is always positive, so the sign of the time derivative of the Lyapunov function is determined only by the size of the coefficient η of the sign function term;

[0042] To satisfy It is always negative, and the size of the coefficient η is selected to satisfy η≥|u2|, which constrains the lower limit of the value of the coefficient η.

[0043] Furthermore, the active blocking control system established in step S3 includes an unknown input sliding mode observer, a current sensing system fault feature extraction module, a voltage sensing system fault feature extraction module, a current controller, a voltage controller, a phase-locked loop module and a sinusoidal pulse width modulation module, wherein the current sensing system fault feature extraction module and the current controller constitute an inner control loop, the voltage sensing system fault feature extraction module and the voltage controller constitute an outer control loop, and the phase-locked loop module is used to make the grid-side current follow the grid-side voltage phase to maintain the unity power factor control function of the single-phase pulse width modulation rectifier; when the sensing system works normally, the measured values ​​of the current sensing system and the voltage sensing system are used to perform closed-loop control on the single-phase pulse width modulation rectifier; when the sensing system fails, the grid-side current observation value or the DC side voltage observation value of the unknown input sliding mode observer is used to replace the measured value of the fault sensing system to perform active blocking control on the single-phase pulse width modulation rectifier, and the control signal is converted by the sinusoidal pulse width modulation module into a switching signal of the switching tube to realize the AC to DC conversion function of the single-phase pulse width modulation rectifier.

[0044] Furthermore, step S3 is specifically as follows:

[0045] The observation errors of the grid-side current and the DC-side voltage are normalized, that is, the normalized values ​​of the state variable observation errors are selected as the residuals, and the residual calculation method of the grid-side current and the DC-side voltage is obtained as follows:

[0046]

[0047] Among them, r is is the grid-side current residual, r udc is the DC side voltage residual, i srmax is the given peak value of the grid-side current, u dcr is the given value of DC side voltage, i s is the grid-side current, u dc is the DC side voltage, is the grid-side current observation value, is the DC side voltage observation value;

[0048] When any sensor system has one of the three faults, the residual of the physical quantity corresponding to the fault sensing will increase, while the residual corresponding to the normal operation sensing will remain below the fault detection threshold or increase in delay. Based on this, the fault characteristics of the single-phase pulse width modulation rectifier are extracted to achieve online detection and location of the fault.

[0049] This method is applicable to three types of sensor system failures: open circuit failure, gain anomaly and noise anomaly.

[0050] Furthermore, the fault detection threshold calculation method of the sensor system is:

[0051]

[0052] Among them, i sth is the detection threshold of the grid-side current sensing system fault, u dcth is the detection threshold of the DC side voltage sensing system fault, e is With e udc are the normalized values ​​of the grid-side current and DC-side voltage observation errors, respectively; n1 and n2 are the current / voltage sensor sampling errors, respectively.

[0053] The system corresponding to the method includes:

[0054] The state variable observation unit is used to construct an unknown input sliding mode observer and observe the state variables online; including: establishing the state equation of the single-phase pulse width modulation rectifier, using symbolic functions to design the parameter matrix, and constructing the unknown input sliding mode observer to realize the online observation of the state variables;

[0055] The unknown input variable estimation and parameter setting unit is used to design the unknown input variable estimation scheme and set the parameters of the unknown input sliding mode observer. It includes: designing an online estimation method for the unknown input variables of the single-phase pulse width modulation rectifier, constructing the Lyapunov function based on the quadratic form of the state variable observation error, and obtaining the lower limit of the coefficient of the symbolic function term of the unknown input sliding mode observer by scaling its time derivative;

[0056] The fault feature extraction and active blocking control unit is used to extract the fault features of the single-phase pulse width modulation rectifier and establish an active blocking control system; it includes: selecting the normalized value of the state variable observation error as the residual, extracting the fault features of the single-phase pulse width modulation rectifier, realizing online detection and positioning of the fault sensing system, and using the observation value of the unknown input sliding mode observer to replace the measured value of the fault sensing system to realize active blocking control until the fault is eliminated.

[0057] An electronic device for storing and executing the method, the device comprising:

[0058] A memory storing executable program code;

[0059] a processor coupled to the memory;

[0060] The processor calls the executable program code stored in the memory to execute the steps of the single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method.

[0061] A computer-readable storage medium for storing and executing the method, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the steps of the single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method.

[0062] Beneficial effects: Compared with the prior art, the significant advantages of the present invention are: (1) there is no need to introduce redundant sensors, and there is no need to increase the cost of software and hardware. It can achieve accurate and rapid extraction of fault characteristics and effective implementation of active blocking control under conditions with unknown inputs, thereby improving the emergency survivability of the single-phase pulse width modulation rectifier under various sensor system failure conditions; (2) it has a simple structure and convenient parameter setting. It does not require complex matrix transformation and iterative calculations, and is more convenient to execute in digital microprocessors commonly used in industrial single-phase pulse width modulation rectifier control. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is a flow chart of the method of the present invention;

[0064] Figure 2 It is the single-phase pulse width modulation rectifier circuit topology of the present invention;

[0065] Figure 3 It is a structural block diagram of the unknown input sliding mode observer of the present invention;

[0066] Figure 4 It is a block diagram of the fault active blocking control structure of the single-phase pulse width modulation rectifier sensing system of the present invention;

[0067] Figure 5 The simulation experiment results of the present invention when the grid-side current sensing system has an open circuit fault, wherein (a) is the simulation result of the observed current and the collected current, (b) is the simulation result of the current residual and the fault sign, and (c) is the simulation result of the actual grid-side current / voltage and the DC-side voltage;

[0068] Figure 6 These are the simulation experiment results of the present invention when the gain of the DC side voltage sensing system is abnormal, wherein (a) is the simulation result of the observed voltage and the collected voltage, (b) is the simulation result of the voltage residual and the fault sign, and (c) is the simulation result of the actual grid side current / voltage and the DC side voltage. DETAILED DESCRIPTION

[0069] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] like Figure 1 As shown, a single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method of the present invention comprises the following steps:

[0071] S1. Constructing an unknown input sliding mode observer to observe state variables online; including: using Kirchhoff's voltage / current law to establish the state equation of a single-phase pulse width modulation rectifier, using symbolic functions to design parameter matrices, and constructing an unknown input sliding mode observer to achieve online observation of state variables; specifically:

[0072] Single-phase pulse width modulation rectifier circuit topology such as Figure 2 As shown. Among them, R is the grid side resistance; L is the grid side inductance; C0 is the DC side voltage stabilizing capacitor; L1 is the DC side filter inductor, C1 is the DC side filter capacitor, L1 and C1 are connected in series to form a DC side filter circuit; u s is the grid-side voltage; i s is the grid-side current; i L is the load current; u dc is the DC side voltage; S1, S2, S3, S4 are the switching functions of the four switching tubes; S A With S B is the equivalent switching function of the two bridge arms. Figure 2 , using Kirchhoff's voltage / current law, the dynamic equation of the single-phase pulse width modulation rectifier is:

[0073]

[0074] Wherein, C is the equivalent capacitance on the DC side.

[0075] Depend on Figure 2 From formula (1), we can see that the two controlled quantities of the single-phase pulse width modulation rectifier are the grid-side current i s With DC side voltage u dc , and there is an unmeasurable input quantity, namely the load current i L , which is defined as an unknown input variable in the present invention.

[0076] Define x as the state variable; y as the output variable; u1 as the known input variable; u2 as the unknown input variable; n×n matrix A as the parameter matrix, n is the number of state variables; n×m1 matrix B1 is the parameter matrix corresponding to the known input variables, m1 is the number of input variable columns; n×m2 matrix B2 is the parameter matrix corresponding to the unknown input variables, m2 is the number of unknown input variable columns; C is the output parameter matrix. Equation (1) can be expressed as a state equation form:

[0077]

[0078] in, is the time derivative of the state variable.

[0079] It can be proved that when the switch function (S A -S B ) takes the value of ±1, the state variables can be observed, that is, there exists a gain matrix H and symmetric positive definite matrices P and Q (Q is taken as the n-order unit matrix), so that the following equation holds:

[0080] (A-HC) T P+P(A-HC)=-Q (3)

[0081] Where T represents the matrix transpose.

[0082] Therefore, there exists a coefficient matrix F such that:

[0083]

[0084] For equation (2), the state variable observation error vector is defined as in is the observed value of the state variable.

[0085] Define the nonlinear feedback matrix E(e) to satisfy:

[0086]

[0087] Among them, η is the coefficient to be designed, η≥ρ>0.

[0088] The nonlinear feedback matrix E(e) is expressed in the form of a symbolic function, ie, E(e) = ηsgn(FCe).

[0089] The product of the output parameter matrix C and the state variable observation error vector e can also be expressed as:

[0090]

[0091] in, is the observed value of the output variable.

[0092] The nonlinear feedback matrix E(e) is transformed into a function of the output variable and its observed value:

[0093]

[0094] Therefore, the unknown input sliding mode observer of the present invention is designed as:

[0095]

[0096] in, is the time derivative of the state variable observation, and G is the gain matrix of the output error feedback term.

[0097] According to formula (8), the state variable x is observed online.

[0098] S2. Design an estimation scheme for unknown input variables of a single-phase pulse width modulation rectifier and perform parameter tuning for an unknown input sliding mode observer; including: designing an online estimation method for unknown input variables of a single-phase pulse width modulation rectifier, constructing a Lyapunov function based on the quadratic form of the state variable observation error, and obtaining the lower limit of the coefficient of the symbolic function term of the unknown input sliding mode observer by scaling its time derivative; specifically:

[0099] According to equations (7) and (8), the estimation method of the unknown input variables of the single-phase pulse width modulation rectifier is:

[0100]

[0101] in, are the observed values ​​of the unknown input variables.

[0102] In order to reconstruct the unknown input variables according to equation (9), a low-pass filter is used to filter them.

[0103] The dynamic equation of the state variable observation error vector e is expressed as:

[0104]

[0105] in, is the time derivative of the state variable observation error, is the time derivative of the state variable.

[0106] The observer parameters are adjusted by a method based on the Lyapunov function. The present invention selects the Lyapunov function V as:

[0107]

[0108] According to equations (3) and (8), the time derivative of the selected Lyapunov function is for:

[0109]

[0110] Since the single-phase pulse width modulation rectifier has only one unknown input variable, namely the load current, the vector u2 is simplified to a variable u2.

[0111] Define the variable σ so that it satisfies The time derivative of the Lyapunov function is further simplified to:

[0112]

[0113] Due to e T The Qe term is always positive, so the sign of the time derivative of the Lyapunov function is determined only by the size of the coefficient η of the sign function term.

[0114] In order to satisfy the requirement that formula (13) is always negative, the size of the coefficient η should be selected so that it satisfies η ≥ |u2|, that is, the lower limit of the value of the coefficient η is constrained.

[0115] The complete structural diagram of the unknown input sliding mode observer of the present invention is as follows: Figure 3As shown, LPF is a low-pass filter. The observer compensates the influence of unknown input variables on the single-phase pulse width modulation rectifier through the switching function term, and realizes accurate observation of state variables through linear feedback. The design process does not require matrix transformation, the method structure is simple, and it is easier to implement digitally.

[0116] S3, extracting the fault characteristics of the single-phase pulse width modulation rectifier and establishing an active blocking control system; including: selecting the normalized value of the state variable observation error as the residual, extracting the fault characteristics of the single-phase pulse width modulation rectifier, realizing online detection and positioning of the fault sensing system, and using the observation value of the unknown input sliding mode observer to replace the measured value of the fault sensing system to realize active blocking control until the fault is eliminated; specifically:

[0117] The present invention considers three types of faults of the grid-side current and DC-side voltage sensing systems, namely, open circuit fault, gain anomaly, and noise anomaly.

[0118] When the load resistance on the DC side of the single-phase pulse width modulation rectifier changes, the grid-side current and DC-side voltage will also change accordingly. Therefore, it is necessary to normalize the observation errors of the grid-side current and DC-side voltage, that is, select the normalized value of the observation error as the residual, and obtain the residual calculation method of the grid-side current and DC-side voltage as follows:

[0119]

[0120] Among them, i srmax is the given peak value of the grid-side current, u dcr is the given value of DC side voltage, r is is the grid-side current residual, r udc is the DC side voltage residual, is the grid-side current observation value, is the observed value of DC side voltage.

[0121] When the sensing system operates normally, the residual error remains below the fault detection threshold.

[0122] When any of the three faults occurs in any sensing system, the residual of the physical quantity corresponding to the fault sensing will increase, while the residual corresponding to the normal operating sensing will remain below the fault detection threshold or the delay will increase. Based on this, the fault characteristics of the single-phase pulse width modulation rectifier can be extracted to achieve online detection and positioning of the fault.

[0123] In the selection of the detection threshold, the observer observation error and the sensor sampling error should be comprehensively considered, and a 3-fold margin should be reserved. Therefore, the detection threshold calculation method of the current / voltage sensor system fault designed by the present invention is as follows:

[0124]

[0125] Among them, i sth is the detection threshold of the grid-side current sensing system fault, u dcth is the detection threshold of the DC side voltage sensing system fault, e is With e udc are the normalized values ​​of the grid-side current and DC-side voltage observation errors, respectively; n1 and n2 are the current / voltage sensor sampling errors, respectively.

[0126] The method of the present invention is applicable to three sensor system fault types: open circuit fault, gain abnormality and noise abnormality.

[0127] When the faulty sensing system is detected and located, an active blocking control strategy is implemented.

[0128] The present invention designs a single-phase pulse width modulation rectifier sensor system fault active blocking control structure block diagram as shown in Figure 4 As shown in the figure, after the sensor system fails, the observed value is used to replace the measured value of the fault sensor system to achieve active blocking control. For example, when the grid-side current sensor system fails, the unknown input sliding mode observer of the grid-side current still works normally, so the grid-side current observed value is used to replace the collected value of the fault current sensor system for closed-loop control until the fault is eliminated; active blocking control can be performed in the same way when the DC side voltage sensor system fails.

[0129] The active blocking control system includes an unknown input sliding mode observer, a current sensing system fault feature extraction module, a voltage sensing system fault feature extraction module, a current controller, a voltage controller, a phase-locked loop (PLL) module, and a sinusoidal pulse width modulation (SPWM) module. Among them, the current sensing system fault feature extraction module and the current controller constitute the inner control loop, the voltage sensing system fault feature extraction module and the voltage controller constitute the outer control loop, and the phase-locked loop module is used to make the grid-side current follow the grid-side voltage phase to maintain the unity power factor control function of the single-phase pulse width modulation rectifier; when the sensing system works normally, the single-phase pulse width modulation rectifier is closed-loop controlled by using the measured values ​​of the current sensing system and the voltage sensing system; when the sensing system fails, the grid-side current observation value or the DC-side voltage observation value of the unknown input sliding mode observer is used to replace the measured value of the fault sensing system to actively block the single-phase pulse width modulation rectifier, and the control signal is converted by the sinusoidal pulse width modulation module into the switching signal of the switch tube to realize the AC to DC conversion function of the single-phase pulse width modulation rectifier.

[0130] The control method of the present invention overcomes the shortcomings of the existing methods such as high complexity, poor portability, and reliance on hardware redundancy, and improves the emergency survivability of the single-phase pulse width modulation rectifier under various sensor system failure conditions.

[0131] The control system used in the above control method includes:

[0132] The state variable observation unit is used to construct an unknown input sliding mode observer and observe the state variables online. Specifically, Kirchhoff's voltage / current law is used to establish the state equation of the single-phase pulse width modulation rectifier, and the symbolic function is used to design the parameter matrix to construct the unknown input sliding mode observer to realize the online observation of the state variables.

[0133] The unknown input variable estimation and parameter setting unit is used to design the unknown input variable estimation scheme and set the parameters of the unknown input sliding mode observer. Specifically, an online estimation method for the unknown input variables of the single-phase pulse width modulation rectifier is designed, and the Lyapunov function is constructed based on the quadratic form of the state variable observation error. The lower limit of the coefficient of the switching function term of the unknown input sliding mode observer is obtained by scaling its time derivative.

[0134] The fault feature extraction and active blocking control unit is used to extract the fault features of the single-phase pulse width modulation rectifier and establish an active blocking control system; specifically: the normalized value of the state variable observation error is selected as the residual, the fault features of the single-phase pulse width modulation rectifier are extracted, the online detection and positioning of the fault sensing system are realized, and the observation value of the unknown input sliding mode observer is used to replace the measured value of the fault sensing system to realize active blocking control until the fault is eliminated.

[0135] An electronic device of the present invention comprises:

[0136] A memory storing executable program code;

[0137] a processor coupled to the memory;

[0138] The processor calls the executable program code stored in the memory to execute the steps of the above-mentioned single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method.

[0139] A computer-readable storage medium of the present invention stores computer instructions, which, when called, are used to execute the steps of the single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method as described above.

[0140] The effectiveness of the fault feature extraction and active blocking control method of the single-phase pulse width modulation rectifier sensor system proposed in the present invention is verified by MATLAB / Simulink simulation experiments and the simulation experimental results are given;

[0141] In order to more clearly demonstrate the proposed control method, the parameters of the high-power single-phase pulse width modulation rectifier selected in the simulation experiment are as follows:

[0142] The grid-side inductance L is 2.3mH, the grid-side resistance R is 0.068Ω, the DC-side voltage stabilizing capacitor C0 is 2200μF, the load resistance is 15Ω, the DC-side filter inductance L1 is 1.15mH, the DC-side filter capacitor C1 is 2200μF, and the grid-side voltage u s The effective value is 1550V, and the DC side voltage is given by u dcr The voltage is 3000V and the modulation frequency is 800Hz.

[0143] In the first group of simulation experiments, an open circuit fault of the grid-side current sensor system was simulated at t = 0.2s, that is, the collected value of the fault sensor system was forced to be set to 0. The simulation results are as follows: Figure 5 In the second group of simulation experiments, the DC side voltage sensor system gain abnormality was simulated at t = 0.2s, and its gain changed from 1 to 1.5. The simulation results are shown in Figure 6 As shown in (a), (b), and (c). It can be seen from the simulation experiment results that the control method proposed in the present invention locates the fault sensing system within 10 sampling cycles after the fault occurs, and at the same time replaces the fault sensing acquisition value with the observed value of the grid-side current or the DC-side voltage for active blocking control. Due to the rapid control reconstruction, the single-phase pulse width modulation rectifier is not affected after the sensor system fails and still works normally. The simulation experiment results confirm that the method proposed in the present invention can quickly locate the fault sensing system and implement active blocking control while realizing accurate observation of the grid-side current and the DC-side voltage, thereby realizing reliable control of the single-phase pulse width modulation rectifier under the condition of sensor loss.

Claims

1. A method for fault feature extraction and active blocking control of a single-phase pulse width modulation rectifier sensor system. It is characterized in that The following steps are involved: S1. Constructing an unknown input sliding mode observer to observe state variables online; including: establishing a state equation of a single-phase pulse width modulation rectifier, using a symbolic function to design a parameter matrix, and constructing an unknown input sliding mode observer to achieve online observation of state variables; S2. Design an estimation scheme for unknown input variables and perform parameter tuning for the unknown input sliding mode observer; including: designing an online estimation method for unknown input variables of a single-phase pulse width modulation rectifier, constructing a Lyapunov function based on the quadratic form of the state variable observation error, and obtaining the lower limit of the coefficient of the symbolic function term of the unknown input sliding mode observer by scaling its time derivative; S3. Extract the fault characteristics of the single-phase pulse width modulation rectifier and establish an active blocking control system; including: selecting the normalized value of the state variable observation error as the residual, extracting the fault characteristics of the single-phase pulse width modulation rectifier, realizing online detection and positioning of the fault sensing system, and using the observation value of the unknown input sliding mode observer to replace the measurement value of the fault sensing system to realize active blocking control until the fault is eliminated.

2. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 1, It is characterized in that In step S1, the state equation of the single-phase pulse width modulation rectifier is established, which is specifically: The dynamic equation of the single-phase pulse width modulation rectifier is established as: Among them, R is the grid side resistance, L is the grid side inductance, u s is the grid-side voltage, i s is the grid-side current, i L is the load current, u dc is the DC side voltage, S A With S B is the equivalent switching function of the two bridge arms, C is the equivalent capacitance on the DC side; The established state equation is: in, is the time derivative of the state variable, x is the state variable, y is the output variable, u 1 is a known input variable, u 2 is the unknown input variable, A is the parameter matrix, B 1 is the parameter matrix corresponding to the known input variables, B 2 is the parameter matrix corresponding to the unknown input variables, and C is the output parameter matrix.

3. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 1, It is characterized in that In step S1, the parameter matrix is ​​designed by using the symbolic function to construct the unknown input sliding mode observer to realize the online observation of the state variables, specifically: Define the state variable observation error vector as in is the observed value of the state variable, and x is the state variable; Define the nonlinear feedback matrix E(e) to satisfy: Among them, η is the coefficient to be designed, η≥ρ>0, ρ is a positive number, F is the coefficient matrix, and C is the output parameter matrix; The nonlinear feedback matrix E(e) is expressed in the form of a symbolic function and transformed into a function of the output variable and its observed value: Among them, y is the output variable, is the observed value of the output variable; The unknown input sliding mode observer is designed as: in, is the time derivative of the observed value of the state variable, u 1 is the known input variable, A is the parameter matrix, B 1 is the parameter matrix corresponding to the known input variables, B 2 is the parameter matrix corresponding to the unknown input variable, G is the gain matrix of the output error feedback term; The state variable x is observed online based on the designed unknown input sliding mode observer.

4. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 1, It is characterized in that Step S2 is specifically as follows: The estimation method for unknown input variables in the design of a single-phase pulse width modulation rectifier is: in, is the observed value of the unknown input variable, η is the coefficient to be designed, η≥ρ>0, ρ is a positive number, F is the coefficient matrix, y is the output variable, is the observed value of the output variable; in order to reconstruct the unknown input variable, a low-pass filter is used to filter it; The dynamic equation of the state variable observation error vector e is expressed as: in, is the time derivative of the state variable observation error, is the time derivative of the observed value of the state variable, is the time derivative of the state variable, A is the parameter matrix, B 2 is the parameter matrix corresponding to the unknown input variable, C is the output parameter matrix, H is the gain matrix, E(e) is the nonlinear feedback matrix, u 2 for unknown input variables; When the state variables are observable, there exists a gain matrix H and symmetric positive definite matrices P and Q such that (A-HC) T P+P(A-HC)=-Q holds; where T represents matrix transpose; The observer parameters are adjusted using the method based on the Lyapunov function. The Lyapunov function V is: V=e T On The vector form u 2 Simplified to a variable u 2 , and define the variable σ to satisfy The time derivative of the Lyapunov function is obtained and simplified to: Among them, e T The Qe term is always positive, so the sign of the time derivative of the Lyapunov function is determined only by the size of the coefficient η of the sign function term; To satisfy is always negative, and the coefficient η is selected to satisfy η≥|u 2 |, constrains the lower limit of the coefficient η.

5. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 1, It is characterized in that The active blocking control system established in step S3 includes an unknown input sliding mode observer, a current sensing system fault feature extraction module, a voltage sensing system fault feature extraction module, a current controller, a voltage controller, a phase-locked loop module and a sinusoidal pulse width modulation module, wherein the current sensing system fault feature extraction module and the current controller constitute an inner control loop, the voltage sensing system fault feature extraction module and the voltage controller constitute an outer control loop, and the phase-locked loop module is used to make the grid-side current follow the grid-side voltage phase to maintain the unity power factor control function of the single-phase pulse width modulation rectifier; when the sensing system works normally, the single-phase pulse width modulation rectifier is closed-loop controlled by using the measured values ​​of the current sensing system and the voltage sensing system; when the sensing system fails, the grid-side current observation value or the DC-side voltage observation value of the unknown input sliding mode observer is used to replace the measured value of the fault sensing system to actively block the single-phase pulse width modulation rectifier, and the control signal is converted by the sinusoidal pulse width modulation module into a switching signal of the switching tube to realize the AC to DC conversion function of the single-phase pulse width modulation rectifier.

6. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 1, It is characterized in that Step S3 is specifically as follows: The observation errors of the grid-side current and the DC-side voltage are normalized, that is, the normalized values ​​of the state variable observation errors are selected as the residuals, and the residual calculation method of the grid-side current and the DC-side voltage is obtained as follows: Among them, r is is the grid-side current residual, r udc is the DC side voltage residual, i srmax is the given peak value of the grid-side current, u dcr is the given value of DC side voltage, i s is the grid-side current, u dc is the DC side voltage, is the observed value of grid-side current, is the DC side voltage observation value; When any sensor system has one of the three faults, the residual of the physical quantity corresponding to the fault sensing will increase, while the residual corresponding to the normal operation sensing will remain below the fault detection threshold or increase in delay. Based on this, the fault characteristics of the single-phase pulse width modulation rectifier are extracted to achieve online detection and location of the fault. This method is applicable to three types of sensor system failures: open circuit failure, gain anomaly and noise anomaly.

7. The single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method according to claim 6, It is characterized in that The fault detection threshold calculation method of the sensing system is: Among them, i sth is the detection threshold of the grid-side current sensing system fault, u dcth is the detection threshold of the DC side voltage sensing system fault, e is With e udc are the normalized values ​​of the grid-side current and DC-side voltage observation errors, n 1 With n 2 are the current / voltage sensor sampling errors respectively.

8. A single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control system, It is characterized in that include: The state variable observation unit is used to construct an unknown input sliding mode observer and observe the state variables online; including: establishing the state equation of the single-phase pulse width modulation rectifier, using symbolic functions to design the parameter matrix, and constructing the unknown input sliding mode observer to realize the online observation of the state variables; The unknown input variable estimation and parameter setting unit is used to design the unknown input variable estimation scheme and set the parameters of the unknown input sliding mode observer. It includes: designing an online estimation method for the unknown input variables of the single-phase pulse width modulation rectifier, constructing the Lyapunov function based on the quadratic form of the state variable observation error, and obtaining the lower limit of the coefficient of the symbolic function term of the unknown input sliding mode observer by scaling its time derivative; The fault feature extraction and active blocking control unit is used to extract the fault features of the single-phase pulse width modulation rectifier and establish an active blocking control system; it includes: selecting the normalized value of the state variable observation error as the residual, extracting the fault features of the single-phase pulse width modulation rectifier, realizing online detection and positioning of the fault sensing system, and using the observation value of the unknown input sliding mode observer to replace the measured value of the fault sensing system to realize active blocking control until the fault is eliminated.

9. An electronic device, It is characterized in that The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the steps of the single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method described in any one of claims 1-7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, which, when called, are used to execute the steps of the single-phase pulse width modulation rectifier sensor system fault feature extraction and active blocking control method described in any one of claims 1-7.

Images