An inverter fault detection method and device based on model parameter identification

Abstract

The application provides an inverter fault detection method and device based on model parameter identification, the method comprising: converting a sampling current of an inverter to form a to-be-detected equivalent current; constructing a first inverter parameter model based on the relationship between the to-be-detected equivalent current and a corresponding voltage; calculating model parameters in the first inverter parameter model based on a fast recursive algorithm to obtain a to-be-detected vector; calculating a fault diagnosis vector corresponding to different open-circuit fault states of the inverter and a standard diagnosis vector corresponding to a normal operation state of the inverter respectively to form a basic matrix; when the to-be-detected vector meets an abnormal condition, comparing the to-be-detected vector with the fault diagnosis vectors in the basic matrix to determine a target fault diagnosis vector; and determining a switch that has an open-circuit fault based on the target fault diagnosis vector. The inverter fault detection method based on model parameter identification can detect a fault switch of the inverter quickly and accurately without being disturbed by load changes of the inverter.

Description

Technical Field

[0001] This invention belongs to the field of inverter fault detection technology, specifically relating to an inverter fault detection method and device based on model parameter identification. Background Technology

[0002] Three-phase voltage source inverters convert electrical energy from DC power sources (including batteries, solar panels, or fuel cells) into AC voltages with variable frequency and amplitude. Due to their high efficiency and flexible control, three-phase voltage source inverters play a crucial role in modern industry and the energy sector. Commonly used power semiconductor switching devices in three-phase voltage source inverters include metal-oxide-semiconductor field-effect transistors (MOSFETs) and insulated-gate bipolar transistors (IGBTs). Recent advancements in power device manufacturing technology have led to the widespread use of wide-bandgap materials such as silicon carbide (SiC) or gallium nitride (GaN) in inverters. Switching faults are the most common problem in inverter systems because the voltage conversion process of a three-phase inverter relies on relatively fragile power semiconductor switching devices that withstand high-voltage stress. Open-circuit faults in inverters can cause distortion of output current and voltage waveforms, abnormal load operation, and overheating problems; therefore, open-circuit fault diagnosis is crucial for ensuring stable inverter operation.

[0003] In voltage-based methods, the actual voltage can be obtained through additional hardware circuitry or voltage sensors. To reduce the need for additional hardware, a hybrid logic dynamic model or a switching function model can be established to derive the required voltage. However, existing voltage detection-based methods always have drawbacks, such as being limited to diagnosing single switch open-circuit faults, requiring improved reliability if the system load is a motor-type load, or having limited adaptability to different loads and poor reliability, etc.

[0004] Open-circuit fault diagnosis methods based on current characteristics exhibit high sensitivity, real-time performance, reliability, and a certain degree of robustness, effectively detecting and diagnosing open-circuit faults in three-phase inverters. However, existing current-characteristic-based detection methods always have drawbacks, such as poor ability to handle load changes, poor detection capability, significant susceptibility to load fluctuations, or the ability to handle only a single open-circuit fault.

[0005] In recent years, there has been considerable research on model-based methods for fault diagnosis of switching transistors in inverter drive systems. Some researchers have proposed a method based on calculated average inter-arm voltage and adaptive error thresholds to diagnose open-circuit faults. This method demonstrates good diagnostic capabilities for various faults; however, it is computationally demanding and does not consider different load conditions. Other researchers have proposed a novel fault feature extraction method based on the trend relationship between adjacent piecewise linear curves. This method can be used to extract fault features and is unaffected by asymmetric interference. This method has a fast fault diagnosis response time; however, it has limitations in handling the number of fault types and requires strong computational support. Still other researchers have established output phase voltage models based on vector decomposition principles and voltage quadratic balance theory, considering neutral point voltage imbalance and time offset. By establishing two phase voltage models to approximate the real system, faults can be located. The fault diagnosis algorithm of this method is relatively complex, and there is still room for improvement in handling load variations.

[0006] Therefore, all of the above indicate that the existing methods are not very effective in diagnosing open-circuit faults in inverters. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide an inverter fault detection method and device based on model parameter identification that can quickly and accurately detect inverter fault switches without being affected by inverter load changes.

[0008] The present invention includes providing an inverter fault detection method based on model parameter identification, comprising:

[0009] The sampled current of the inverter is converted and processed to form the equivalent current to be measured;

[0010] A first inverter parameter model is constructed based on the relationship between the measured equivalent current and the corresponding voltage.

[0011] The model parameters in the parameter model of the first inverter are calculated and determined based on a fast recursive algorithm to obtain the vector to be detected.

[0012] The fault diagnosis vectors for different open-circuit fault states of the inverter and the standard diagnosis vectors for the normal operation state of the inverter are calculated and determined respectively, and a basic matrix is ​​formed based on the fault diagnosis vectors and the standard diagnosis vectors.

[0013] When the detection vector meets the abnormal conditions based on the standard diagnostic vector, the detection vector is compared with the fault diagnosis vector in the basic matrix, and the target fault diagnosis vector is determined based on the comparison result.

[0014] The switch in the inverter that has an open-circuit fault is determined based on the target fault diagnosis vector.

[0015] In some embodiments, the step of converting the sampled current of the inverter to form the equivalent current to be measured includes:

[0016] The sampled current of the inverter is subjected to Clarke conversion to form the equivalent current to be measured.

[0017] In some embodiments, the method further includes:

[0018] The ratio of the measured equivalent current to the amplitude of the equivalent voltage and equivalent current when the inverter is under rated operating conditions is calculated to normalize the measured equivalent current.

[0019] In some embodiments, constructing the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage includes:

[0020] Based on the relationship between the output current and voltage of each phase in the inverter and the Clarke transform algorithm, the normalized relationship expression between the measured equivalent current and the corresponding voltage is determined.

[0021] The relational expression is discretized to form the parameter model of the first inverter.

[0022] In some embodiments, the step of calculating and determining the model parameters in the first inverter parameter model based on a fast recursive algorithm to obtain the vector to be detected includes:

[0023] The model parameter calculation formula is generated based on the linear expression of the fast recursive algorithm and the discrete expression of the first inverter parameter model:

[0024]

[0025] in, y = i N (k)-i N (k-1), i N The normalized equivalent current to be measured, k is the number of current samples. Both represent the voltage corresponding to the collected equivalent current;

[0026] The model parameters are calculated and determined based on the model parameter calculation formula to obtain the vector to be detected.

[0027] In some embodiments, calculating and determining the fault diagnosis vector corresponding to different open-circuit fault states of the inverter and the standard diagnosis vector corresponding to the normal operating state of the inverter includes:

[0028] Obtain the current of the inverter under different open-circuit fault conditions and normal operating conditions;

[0029] Each of the currents is converted to form an equivalent current;

[0030] Based on the relationship between each equivalent current and the corresponding voltage, a corresponding second inverter parameter model is constructed.

[0031] The model parameters in each second inverter parameter model are calculated and determined based on a fast recursive algorithm, thereby obtaining the fault diagnosis vectors for different open-circuit fault states of the corresponding inverter and the standard diagnosis vectors for the normal operating state of the corresponding inverter.

[0032] In some embodiments, the method further includes:

[0033] Calculate and determine the first Eulerian distance between the vector to be detected and the standard diagnostic vector;

[0034] The detection vector is determined based on the first Eulerian distance to determine whether it meets the abnormal conditions.

[0035] In some embodiments, comparing the vector to be detected with the fault diagnosis vector in the base matrix and determining the target fault diagnosis vector based on the comparison result includes:

[0036] Calculate the second Eulerian distance between the vector to be detected and each fault diagnosis vector in the fundamental matrix;

[0037] The target fault diagnosis vector is determined based on multiple second Eulerian distances.

[0038] In some embodiments, at least each fault diagnosis vector in the basic matrix has a unique identifier, which is used to indicate the open-circuit fault condition of the inverter represented by the corresponding diagnosis vector.

[0039] The determination of the target fault diagnosis vector based on multiple second Eulerian distances includes:

[0040] Determine the minimum value among multiple second Eulerian distances, and determine the fault diagnosis vector corresponding to the minimum value as the target fault diagnosis vector;

[0041] The step of determining the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector includes:

[0042] The switch in the inverter that has an open-circuit fault can be directly determined based on the unique identifier of the target fault diagnosis vector, or the switch in the inverter that has an open-circuit fault can be determined based on the stored fault information that matches the unique identifier.

[0043] Another embodiment of the present invention also provides an inverter fault detection device based on model parameter identification, comprising:

[0044] The conversion module is used to convert the sampled current of the inverter into the equivalent current to be measured;

[0045] The construction module is used to construct the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage;

[0046] The first calculation module is used to calculate and determine the model parameters in the parameter model of the first inverter according to the fast recursive algorithm, and obtain the vector to be detected.

[0047] The second calculation module is used to calculate and determine the fault diagnosis vector for different open-circuit fault states of the corresponding inverter and the standard diagnosis vector for the normal operation state of the corresponding inverter, and form a basic matrix based on the fault diagnosis vector and the standard diagnosis vector.

[0048] The first determining module is used to compare the vector to be detected with the fault diagnosis vector in the basic matrix when the standard diagnostic vector determines that the vector to be detected meets the abnormal conditions, and to determine the target fault diagnosis vector based on the comparison result.

[0049] The second determining module is used to determine the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector.

[0050] The beneficial effects of this invention are that it combines a parametric model of the three-phase inverter current with a fast recursive algorithm to calculate vectors that characterize the current characteristics of the inverter under different open-circuit fault conditions. Based on the comparison between these vectors, the switch in the inverter where the open-circuit fault occurs is determined. The method is simple, low-cost, more sensitive to fault detection, more accurate in fault location, and has a faster diagnostic speed compared to traditional current diagnostic methods. The method based on this embodiment can evaluate model parameters in real time without adding extra hardware; it can be implemented using only existing equipment. Attached Figure Description

[0051] Figure 1 This is a flowchart of an inverter fault detection method based on model parameter identification in an embodiment of the present invention.

[0052] Figure 2 This is a circuit diagram of the inverter in an embodiment of the present invention.

[0053] Figure 3 This is an equivalent circuit diagram of the inverter circuit load in an embodiment of the present invention.

[0054] Figure 4 The diagram shows the current waveforms when different faults occur in the inverters according to embodiments of the present invention.

[0055] Figure 5 This is a diagram illustrating the fault diagnosis vector formation mechanism in an embodiment of the present invention.

[0056] Figure 6 This is a flowchart illustrating the application of the fault detection switch in this embodiment of the invention.

[0057] Figure 7 This is a structural block diagram of the inverter fault detection device based on model parameter identification in an embodiment of the present invention. Detailed Implementation

[0058] like Figure 1 As shown, the present invention includes an inverter fault detection method based on model parameter identification, comprising:

[0059] S1: Convert the sampled current of the inverter to form the equivalent current to be measured;

[0060] S2: Construct the parameter model of the first inverter based on the relationship between the measured equivalent current and the corresponding voltage;

[0061] S3: Calculate and determine the model parameters in the parameter model of the first inverter based on the fast recursive algorithm to obtain the vector to be detected;

[0062] S4: Calculate and determine the fault diagnosis vector for different open-circuit fault states of the corresponding inverter and the standard diagnosis vector for the normal operation state of the corresponding inverter, and form a basic matrix based on the fault diagnosis vector and the standard diagnosis vector.

[0063] S5: When the detected vector meets the abnormal conditions based on the standard diagnostic vector, the detected vector is compared with the fault diagnosis vector in the basic matrix, and the target fault diagnosis vector is determined based on the comparison result.

[0064] S6: Determine the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector.

[0065] Based on the above, this embodiment combines a parameter model of the three-phase inverter current with a fast recursive algorithm to calculate vectors that characterize the current characteristics of the inverter under different open-circuit fault conditions. Based on the comparison between these vectors, the switch in the inverter experiencing the open-circuit fault is determined. The method is simple, low-cost, more sensitive to fault detection, more accurate in fault location, and has a faster diagnostic speed compared to traditional current diagnostic methods. The method in this embodiment can evaluate model parameters in real time without requiring additional hardware; it can be implemented using existing equipment. Furthermore, the detection method in this embodiment has a wide range of applications, suitable not only for general load systems but also for various application scenarios such as traction drive systems.

[0066] Specifically, such as Figure 2As shown, the inverter in this embodiment has six identical IGBT switches T1-T6, and D1-D6 are anti-parallel diodes corresponding to each switch. Gate signals g1-g6 control the switching on and off of switches T1-T6. The gate signals of the two switches on the same bridge arm of the inverter are complementary; when the gate signal is 1, the corresponding switch is on, and when the gate signal is 0, the switch is off. The inverter in this embodiment uses common modulation techniques, including sinusoidal pulse width modulation (SPWM) to generate the gate signal. As shown in the figure, the load of the three-phase inverter in this embodiment is a star-connected RL load, and typically an LC filter to suppress higher-order harmonics is connected to its AC side. A L B L C C A C B C C These are the inductors and capacitors in the inverter's filter circuit.

[0067] When the inverter is in normal operation, the three-phase output current is a sine wave with equal amplitude in each phase and a phase difference of 120°. The expression for the current is:

[0068]

[0069] Among them I Am I Bm I Cm These are the amplitudes of the three-phase currents, which are normally equal. ω is the current angular frequency, and θ is the phase difference.

[0070] According to the principle of circuit equivalence Figure 2 The load circuit of the inverter shown can be simplified to a three-phase equivalent RLC circuit per phase, as detailed below. Figure 3 As shown. Where j = a, b, c, U j I j Represents the three-phase input voltage and output current, R eq L eq C eq These are the corresponding equivalent reactances. According to Kirchhoff's current law, the relationship between the output current and voltage of each phase can be:

[0071] i j =i Req +i Leq +i Ceq

[0072]

[0073] Furthermore, in three-phase voltage-source inverters, the most common open-circuit fault is a single-switch or double-switch open-circuit fault. The inverter in this embodiment has a total of 22 scenarios, including normal operation and 21 fault scenarios. Of course, the specific scenarios are not unique; the number and types of faults may vary for different inverters. In this embodiment, fault types can be divided into the following three categories. To facilitate identification of each fault, this embodiment assigns a numerical label to each fault, as follows:

[0074] Fault type 1 (single tube fault, fault labels 1-6): T1, T2, T3, T4, T5, T6.

[0075] Fault type 2 (in-phase dual-tube fault, fault labels 7-9): T1&T2, T3&T4, T5&T6.

[0076] Fault Type 3 (Anisotropic Dual-Tube Fault, Fault Label 10-21): T1&T3, T1&T4, T1&T5, T1&T6, T2&T3, T2&T4, T2&T5, T2&T6, T3&T5, T3&T6, T4&T5, T4&T6.

[0077] normal 0 <![CDATA[T1&T4]]> 11 <![CDATA[T1]]> 1 <![CDATA[T1&T5]]> 12 <![CDATA[T2]]> 2 <![CDATA[T1&T6]]> 13 <![CDATA[T3]]> 3 <![CDATA[T2&T3]]> 14 <![CDATA[T4]]> 4 <![CDATA[T2&T4]]> 15 <![CDATA[T5]]> 5 <![CDATA[T2&T5]]> 16 <![CDATA[T6]]> 6 <![CDATA[T2&T6]]> 17 <![CDATA[T1&T2]]> 7 <![CDATA[T3&T5]]> 18 <![CDATA[T3&T4]]> 8 <![CDATA[T3&T6]]> 19 <![CDATA[T5&T6]]> 9 <![CDATA[T4&T5]]> 20 <![CDATA[T1&T3]]> 10 <![CDATA[T4&T6]]> 21

[0078] Storing the above fault information can be used for subsequent testing to identify the faulty switch.

[0079] In this embodiment, the sampled current of the inverter is converted to form the equivalent current to be measured, including:

[0080] S7: The sampled current of the inverter is processed by Clarke conversion to form the equivalent current to be measured.

[0081] For example, the Clarke transform can convert a three-phase AC voltage or current into two independent voltage and current signals I. α and I β This transformation makes it easier to observe changes in current under various inverter fault conditions. The Clarke transformation formula is as follows, and based on this formula, the sampled current can be transformed to form the equivalent current to be measured:

[0082]

[0083] Under different operating states or fault conditions of the inverter, the inverter's I α and I β Each has its own unique characteristics. For example Figure 4 As shown, it illustrates the I value of the inverter under normal operating conditions. α and I β And the equivalent current state under the other three fault conditions. Figure 4 (a) shows the output waveform of the α-β current under normal conditions, which is a sine wave. Figure 4 (b)-(d) illustrate the different current modes generated by the inverter when an open-circuit fault occurs in the inverter system, each with a unique waveform. Although the frequency of the α-β current remains constant, the amplitude and current value will vary, thus the change in the α-β current can reflect the fault condition of the inverter's switching.

[0084] In some embodiments, the method further includes:

[0085] S8: Calculate the ratio of the measured equivalent current to the amplitude of the equivalent voltage and equivalent current when the inverter is under rated operating conditions, so as to normalize the measured equivalent current.

[0086] Since the α-β current is affected by load variations, it influences the estimated parameter values. To eliminate this effect, the measured α-β current values ​​need to be normalized after the current undergoes a Clarke transform. The normalization formula is as follows:

[0087] I N =I αβ / I αβm

[0088] Among them, I N This is the normalized equivalent current value to be measured. αβm These are the voltage and current amplitudes of α-β under normal rated operating conditions.

[0089] Furthermore, after completing the current processing, it is necessary to construct the first inverter parameter model, which includes:

[0090] S9: Based on the relationship between the output current and voltage of each phase in the inverter and the Clarke transform algorithm, determine the relationship expression between the normalized equivalent current to be measured and the corresponding voltage.

[0091] S10: Discretize the relational expression to form the first inverter parameter model.

[0092] Specifically, based on the above-described relationship between the output current and voltage of each phase, under normal conditions, the relationship between the α-β voltage and the α-β current can be expressed as:

[0093]

[0094] Taking the derivative of both sides of the above equation, we get:

[0095]

[0096] Assume the sampling time is TS Compared to the load cycle time, the sampling time is a very small time, therefore the discrete expression of this formula is:

[0097]

[0098] The above equation can be simplified to:

[0099] i αβ (k)-i αβ (k-1)=Au αβ (k)+Bu αβ (k-1)+Cu αβ (k-2)

[0100] Where A, B, and C are model parameters, represented as follows:

[0101]

[0102]

[0103]

[0104] Therefore, for a three-phase inverter, under normal conditions, the linear relationship between its current and voltage can be written as:

[0105]

[0106] Where Y = i αβ (k)-i αβ (k-1) represents the difference between the two discrete currents, and ε(k) represents the system's white noise.

[0107] Furthermore, the model parameters in the first inverter parameter model are calculated and determined based on a fast recursive algorithm to obtain the detection vector, including:

[0108] S11: Generating model parameter calculation formulas based on the linear expression of the fast recursive algorithm and the discrete expression of the first inverter parameter model:

[0109]

[0110] in, y = i N (k)-i N (k-1), i N The normalized equivalent current to be measured, k is the number of current samples. Both represent the voltage corresponding to the collected equivalent current;

[0111] S12: Calculate and determine the model parameters based on the model parameter calculation formula to obtain the vector to be detected.

[0112] For example, a nonlinear discrete-time dynamic system is set up based on a fast recursive algorithm, and its linear expression is:

[0113] y=ΨΘ+Ξ

[0114] in Let ψ = [u] be the regression moment formed by the voltage corresponding to each sampled current. ref (k)u ref (k-1)u ref [(k-2)], U ref This is the rated reference voltage.

[0115] In the fast recursive algorithm, two recursion matrices M are predefined. k and R k To complete the recursive process, see the following:

[0116]

[0117]

[0118] in The first k columns of the complete regression matrix Ψ are included, where k = 1, ..., S, and R0 = I. Therefore, by selecting the first k columns of Ψ, the parameters of the minimization cost function and the related minimization cost function can be formally estimated as follows:

[0119]

[0120]

[0121] To simplify the formula and reduce computational complexity, the following three quantities are defined to simplify recursion:

[0122]

[0123]

[0124]

[0125] Where j = 1,...,S and k = 1,...,S. According to R k The properties allow for the explicit calculation of the new parameter term φ. k+1 Net contribution to the cost function:

[0126]

[0127] By calculating the net contribution of each item, the model item with the largest contribution is selected one by one. Finally, after selecting all important model items, the parameter value for each selected item is calculated:

[0128]

[0129] like Figure 5 As shown, the fault diagnosis vectors for different open-circuit fault states of the inverter and the standard diagnosis vectors for the normal operating states of the inverter are calculated and determined respectively, including:

[0130] S13: Obtain the current of the inverter under different open-circuit fault conditions and normal operating conditions;

[0131] S14: Convert each current to form an equivalent current;

[0132] S15: Construct the corresponding second inverter parameter model based on the relationship between each equivalent current and the corresponding voltage;

[0133] S16: Calculate and determine the model parameters in the parameter model of each second inverter based on the fast recursive algorithm, and obtain the fault diagnosis vectors for different open circuit fault states of the corresponding inverter and the standard diagnosis vectors for the normal operation state of the corresponding inverter.

[0134] In other words, based on the aforementioned method, the standard diagnostic vector for the inverter under normal operating conditions and the fault diagnostic vector under different open-circuit fault conditions are calculated sequentially. For example, the above operation can be performed using a three-phase inverter system. The three-phase inverter system is treated as a black box, and all fault conditions are simulated. The input to the black box is the three-phase voltage, that is, the inverter reference voltage U with known constant amplitude and phase angle under normal conditions. ref The output of the black box is the measured current signal. The output current signal is sampled and then converted to its equivalent α-β values ​​using the Clarke transform. Parametric model equations are constructed for the α and β currents. After processing by the parametric model equations and the fast regression algorithm, six model parameters A are obtained. α B α C α A β B β C β This forms a vector (A) with 6 elements. α B α C α A β B β C β For each different fault condition listed in Table 1, parameter identification using a fast recursive algorithm yields parameter vectors for the inverter under different scenarios. The 22 different parameter estimation vectors represent the characteristics of each inverter condition and can serve as a reference benchmark for inverter fault diagnosis.

[0135] Each set of measured three-phase current values ​​can be converted into a fault detection vector (FDV) through the above process, that is: FDV = [A α B α C α A β B β C β ] T The basic matrix is ​​constructed based on all diagnostic vectors and standard diagnostic vectors. This matrix is ​​used to detect and locate open-circuit faults in two-level three-phase inverter systems.

[0136] When determining whether the vector to be detected is abnormal, the method further includes:

[0137] S17: Calculate and determine the first Eulerian distance between the vector to be detected and the standard diagnostic vector;

[0138] S18: Determine whether the vector to be detected satisfies the abnormal condition based on the first Eulerian distance.

[0139] For example, the k-nearest neighbors (kNN) algorithm is a simple yet effective classification method that uses proximity to determine whether a dataset is grouped around a certain point. For each new input instance, the algorithm uses distance as a metric to determine the k nearest instances in the training dataset. The input instance is then classified into one of these k nearest neighbor classes.

[0140] In the kNN algorithm, similarity between samples is calculated using distance. This embodiment uses the Euclidean distance of vectors for this determination. Let x = [x1, x2, ..., x...]. n ] T and y = [y1, y2, ..., y n ] T R is a real vector space n The n-dimensional vector in the vector. The Eulerian distance between the two is defined as follows:

[0141]

[0142] After establishing the basic matrix, each fault diagnosis vector (FDV) corresponds to a specific inverter fault state, which can be distinguished through calculation. Applying this method to the fault diagnosis vectors (FDV) in the basic matrix, the Eulerian distance between two fault diagnosis vectors (FDV) can be represented by Vd: Vd = d(FDV) est FDV nomal )

[0143] Among them, FDV estIt is a fault diagnosis vector composed of real-time monitored inverter current, FDV nomal This is the diagnostic vector under normal operating conditions. When the inverter is operating normally under different loads, the inverter current waveform only changes in amplitude. Since the system parameters remain essentially unchanged, the corresponding fault diagnosis vector remains essentially constant. Therefore, Vd changes very little. However, if Vd is large, this indicates a fault, which can be used to detect an open-circuit fault. The diagnostic function is:

[0144]

[0145] Here, K is a gain factor that can be set based on actual observations of the noise level in the estimated parameter set. If Fd = 1, it indicates that the inverter has an open-circuit fault. If Fd = 0, it indicates that the inverter is operating normally.

[0146] like Figure 6 As shown, when an anomaly is determined in the vector to be detected, the vector to be detected is compared with the fault diagnosis vector in the fundamental matrix, and the target fault diagnosis vector is determined based on the comparison result. Specifically, this includes:

[0147] S19: Calculate the second Eulerian distance between the vector to be detected and each fault diagnosis vector in the basis matrix;

[0148] S20: Determine the target fault diagnosis vector based on multiple second Eulerian distances.

[0149] Among them, the target fault diagnosis vector is determined based on multiple second Eulerian distances, including:

[0150] S21: Determine the minimum value among multiple second Eulerian distances, and determine the fault diagnosis vector corresponding to the minimum value as the target fault diagnosis vector;

[0151] That is, the second Eulerian distance value is obtained by comparing the measured FDV data with the other 21 fault diagnosis vectors in the fundamental matrix. According to the proximity principle, the fault tag corresponding to the minimum value indicates an open-circuit fault. In this case, the fault location function is:

[0152] Vd i =d(FDV) est -FDV basei ) i＝1-21

[0153] F location =min(Vd) i ) i＝1-21

[0154] FDV baseiIt is the feature vector corresponding to the fault identified in Table 1 of the basic matrix, which is the target fault diagnosis vector in this embodiment.

[0155] In this embodiment, at least each fault diagnosis vector in the basic matrix has a unique identifier. The unique identifier is used to indicate the open-circuit fault condition of the inverter represented by the corresponding diagnosis vector. For example, the unique identifier is named based on the fault switch, etc., and the specifics are not fixed.

[0156] The switches in the inverter that have experienced open-circuit faults, determined based on the target fault diagnosis vector, include:

[0157] S22: Directly determine the switch in the inverter that has an open circuit fault based on the unique identifier of the target fault diagnosis vector, or determine the switch in the inverter that has an open circuit fault based on the stored fault information that matches the unique identifier.

[0158] For example, if a unique identifier can directly identify the faulty circuit switch, then it can be directly identified. If not, as mentioned earlier, each fault diagnosis vector only has a number label. In this case, the stored data table can be used to match and find the information corresponding to the number label of the target fault diagnosis vector. This information expresses the faulty switch.

[0159] like Figure 7 As shown, another embodiment of the present invention also provides an inverter fault detection device 100 based on model parameter identification, comprising:

[0160] The conversion module is used to convert the sampled current of the inverter into the equivalent current to be measured;

[0161] The construction module is used to construct the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage;

[0162] The first calculation module is used to calculate and determine the model parameters in the parameter model of the first inverter according to the fast recursive algorithm, and obtain the vector to be detected.

[0163] The second calculation module is used to calculate and determine the fault diagnosis vector for different open-circuit fault states of the corresponding inverter and the standard diagnosis vector for the normal operation state of the corresponding inverter, and form a basic matrix based on the fault diagnosis vector and the standard diagnosis vector.

[0164] The first determining module is used to compare the vector to be detected with the fault diagnosis vector in the basic matrix when the standard diagnostic vector determines that the vector to be detected meets the abnormal conditions, and to determine the target fault diagnosis vector based on the comparison result.

[0165] The second determining module is used to determine the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector.

[0166] In some embodiments, the step of converting the sampled current of the inverter to form the equivalent current to be measured includes:

[0167] The sampled current of the inverter is subjected to Clarke conversion to form the equivalent current to be measured.

[0168] In some embodiments, the method further includes:

[0169] The ratio of the measured equivalent current to the amplitude of the equivalent voltage and equivalent current when the inverter is under rated operating conditions is calculated to normalize the measured equivalent current.

[0170] In some embodiments, constructing the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage includes:

[0171] Based on the relationship between the output current and voltage of each phase in the inverter and the Clarke transform algorithm, the normalized relationship expression between the measured equivalent current and the corresponding voltage is determined.

[0172] The relational expression is discretized to form the parameter model of the first inverter.

[0173] In some embodiments, the step of calculating and determining the model parameters in the first inverter parameter model based on a fast recursive algorithm to obtain the vector to be detected includes:

[0174] The model parameter calculation formula is generated based on the linear expression of the fast recursive algorithm and the discrete expression of the first inverter parameter model:

[0175]

[0176] in, y = i N (k)-i N (k-1), i N The normalized equivalent current to be measured, k is the number of current samples. Both represent the voltage corresponding to the collected equivalent current;

[0177] The model parameters are calculated and determined based on the model parameter calculation formula to obtain the vector to be detected.

[0178] In some embodiments, calculating and determining the fault diagnosis vector corresponding to different open-circuit fault states of the inverter and the standard diagnosis vector corresponding to the normal operating state of the inverter includes:

[0179] Obtain the current of the inverter under different open-circuit fault conditions and normal operating conditions;

[0180] Each of the currents is converted to form an equivalent current;

[0181] Based on the relationship between each equivalent current and the corresponding voltage, a corresponding second inverter parameter model is constructed.

[0182] The model parameters in each second inverter parameter model are calculated and determined based on a fast recursive algorithm, thereby obtaining the fault diagnosis vectors for different open-circuit fault states of the corresponding inverter and the standard diagnosis vectors for the normal operating state of the corresponding inverter.

[0183] In some embodiments, the apparatus further includes:

[0184] The third calculation module is used to calculate and determine the first Eulerian distance between the vector to be detected and the standard diagnostic vector;

[0185] The third determining module is used to determine whether the vector to be detected satisfies the abnormal condition based on the first Eulerian distance.

[0186] In some embodiments, comparing the vector to be detected with the fault diagnosis vector in the base matrix and determining the target fault diagnosis vector based on the comparison result includes:

[0187] Calculate the second Eulerian distance between the vector to be detected and each fault diagnosis vector in the fundamental matrix;

[0188] The target fault diagnosis vector is determined based on multiple second Eulerian distances.

[0189] In some embodiments, at least each fault diagnosis vector in the basic matrix has a unique identifier, which is used to indicate the open-circuit fault condition of the inverter represented by the corresponding diagnosis vector.

[0190] The determination of the target fault diagnosis vector based on multiple second Eulerian distances includes:

[0191] Determine the minimum value among multiple second Eulerian distances, and determine the fault diagnosis vector corresponding to the minimum value as the target fault diagnosis vector;

[0192] The step of determining the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector includes:

[0193] The switch in the inverter that has an open-circuit fault can be directly determined based on the unique identifier of the target fault diagnosis vector, or the switch in the inverter that has an open-circuit fault can be determined based on the stored fault information that matches the unique identifier.

[0194] Another embodiment of the present invention also provides an electronic device, comprising:

[0195] At least one processor; and,

[0196] A memory communicatively connected to the at least one processor; wherein,

[0197] The memory stores instructions that can be executed by the at least one processor to implement the inverter fault detection method based on model parameter identification as described in any of the embodiments above.

[0198] Furthermore, one embodiment of the present invention also provides a storage medium storing a computer program, which, when executed by a processor, implements the inverter fault detection method based on model parameter identification as described above. It should be understood that the various solutions in this embodiment have the corresponding technical effects in the above-described method embodiments, and will not be repeated here.

[0199] Furthermore, embodiments of the present invention also provide a computer program product, which is tangibly stored on a computer-readable medium and includes computer-readable instructions that, when executed, cause at least one processor to perform an inverter fault detection method based on model parameter identification as described in the embodiments above.

[0200] It should be noted that the computer storage medium of the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access storage medium (RAM), a read-only storage medium (ROM), an erasable programmable read-only storage medium (EPROM or flash memory), an optical fiber, a portable compact disk read-only storage medium (CD-ROM), an optical storage medium, a magnetic storage medium, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program configured for use by or in connection with an instruction execution system, system, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, antenna, optical fiber, RF, etc., or any suitable combination thereof.

[0201] Furthermore, those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0202] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1One or more processes and / or boxes Figure 1 A system that specifies functions in one or more boxes.

[0203] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including an instruction set implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0204] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0205] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0206] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A fault detection method for inverters based on model parameter identification, characterized in that, include: The sampled current of the inverter is converted and processed to form the equivalent current to be measured; A first inverter parameter model is constructed based on the relationship between the measured equivalent current and the corresponding voltage. The model parameters in the parameter model of the first inverter are calculated and determined based on a fast recursive algorithm to obtain the vector to be detected. The fault diagnosis vectors for different open-circuit fault states of the inverter and the standard diagnosis vectors for the normal operation state of the inverter are calculated and determined respectively, and a basic matrix is ​​formed based on the fault diagnosis vectors and the standard diagnosis vectors. When the detection vector meets the abnormal conditions based on the standard diagnostic vector, the detection vector is compared with the fault diagnosis vector in the basic matrix, and the target fault diagnosis vector is determined based on the comparison result. The switch in the inverter that has an open-circuit fault is determined based on the target fault diagnosis vector; The construction of the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage includes: Based on the relationship between the output current and voltage of each phase in the inverter and the Clarke transform algorithm, the normalized relationship expression between the measured equivalent current and the corresponding voltage is determined. Discretize the relational expression to form the parameter model of the first inverter; The step of calculating and determining the model parameters in the first inverter parameter model based on a fast recursive algorithm to obtain the vector to be detected includes: The model parameter calculation formula is generated based on the linear expression of the fast recursive algorithm and the discrete expression of the first inverter parameter model: in, , , The normalized equivalent current to be measured, , This represents the number of current samples. Both represent the voltage corresponding to the collected equivalent current; The model parameters are calculated and determined based on the model parameter calculation formula to obtain the vector to be detected; The calculation and determination of fault diagnosis vectors for different open-circuit fault states of the inverter and standard diagnosis vectors for the normal operating state of the inverter includes: Obtain the current of the inverter under different open-circuit fault conditions and normal operating conditions; Each of the currents is converted to form an equivalent current; Based on the relationship between each equivalent current and the corresponding voltage, a corresponding second inverter parameter model is constructed. The model parameters in each second inverter parameter model are calculated and determined based on a fast recursive algorithm, thereby obtaining the fault diagnosis vectors for different open-circuit fault states of the corresponding inverter and the standard diagnosis vectors for the normal operating state of the corresponding inverter.

2. The inverter fault detection method based on model parameter identification according to claim 1, characterized in that, The process of converting the sampled current of the inverter to form the equivalent current to be measured includes: The sampled current of the inverter is subjected to Clarke conversion to form the equivalent current to be measured.

3. The inverter fault detection method based on model parameter identification according to claim 2, characterized in that, The method further includes: The ratio of the measured equivalent current to the amplitude of the equivalent voltage and equivalent current when the inverter is under rated operating conditions is calculated to normalize the measured equivalent current.

4. The inverter fault detection method based on model parameter identification according to claim 1, characterized in that, The method further includes: Calculate and determine the first Eulerian distance between the vector to be detected and the standard diagnostic vector; The detection vector is determined based on the first Eulerian distance to determine whether it meets the abnormal conditions.

5. The inverter fault detection method based on model parameter identification according to claim 1, characterized in that, The step of comparing the vector to be detected with the fault diagnosis vector in the basic matrix and determining the target fault diagnosis vector based on the comparison result includes: Calculate the second Eulerian distance between the vector to be detected and each fault diagnosis vector in the fundamental matrix; The target fault diagnosis vector is determined based on multiple second Eulerian distances.

6. The inverter fault detection method based on model parameter identification according to claim 5, characterized in that, At least each fault diagnosis vector in the basic matrix has a unique identifier, which is used to indicate the open-circuit fault condition of the inverter represented by the corresponding diagnosis vector. The determination of the target fault diagnosis vector based on multiple second Eulerian distances includes: Determine the minimum value among multiple second Eulerian distances, and determine the fault diagnosis vector corresponding to the minimum value as the target fault diagnosis vector; The step of determining the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector includes: The switch in the inverter that has an open-circuit fault can be directly determined based on the unique identifier of the target fault diagnosis vector, or the switch in the inverter that has an open-circuit fault can be determined based on the stored fault information that matches the unique identifier.

7. An inverter fault detection device based on model parameter identification, characterized in that, The method for implementing the inverter fault detection method based on model parameter identification as described in any one of claims 1-6 includes: The conversion module is used to convert the sampled current of the inverter into the equivalent current to be measured; The construction module is used to construct the first inverter parameter model based on the relationship between the measured equivalent current and the corresponding voltage; The first calculation module is used to calculate and determine the model parameters in the parameter model of the first inverter according to the fast recursive algorithm, and obtain the vector to be detected. The second calculation module is used to calculate and determine the fault diagnosis vector for different open-circuit fault states of the corresponding inverter and the standard diagnosis vector for the normal operation state of the corresponding inverter, and form a basic matrix based on the fault diagnosis vector and the standard diagnosis vector. The first determining module is used to compare the vector to be detected with the fault diagnosis vector in the basic matrix when the standard diagnostic vector determines that the vector to be detected meets the abnormal conditions, and to determine the target fault diagnosis vector based on the comparison result. The second determining module is used to determine the switch in the inverter that has an open-circuit fault based on the target fault diagnosis vector.

Images