Open-circuit diagnostic method for multiphase motors with combined harmonic spatial current trajectories

Abstract

This invention discloses a multiphase motor open-circuit diagnosis method based on combined harmonic spatial current trajectories, comprising: filtering the multiphase motor winding current and transforming it through spatial coordinates to obtain orthogonal current components of a specific order harmonic subspace; linearly transforming these components to generate current variables corresponding to the number of motor phases, and calculating the results through inversion and moving average; setting a threshold to avoid the influence of single-point trajectories, defining a judgment factor, and determining the fault phase and type by matching it with a preset diagnostic table. This method fully exploits the characteristics of harmonic faults, can flexibly adjust parameters to adapt to multiphase motors with different operating conditions and topologies, effectively solves the problems of insufficient diagnostic accuracy due to reliance on the fundamental frequency, easy misjudgment and omission under complex operating conditions, and poor adaptability, and improves diagnostic accuracy and reliability, making it suitable for multiphase motor open-circuit fault diagnosis scenarios.

Description

Technical Field

[0001] This invention relates to the field of open-circuit diagnostic technology for multiphase motors, and more particularly to a method for diagnosing open-circuit faults in multiphase motors using a combined harmonic spatial current trajectory. Existing technology

[0002] In industrial production, transportation, and other fields, multiphase motors are widely used in equipment requiring high operational stability due to their high power density and strong fault tolerance. However, during long-term operation, multiphase motors are prone to open-circuit faults due to factors such as winding aging, insulation damage, and loose wiring. If such faults are not detected and located in a timely manner, they can lead to decreased motor efficiency, increased energy consumption, and even problems such as overheating of motor windings and increased torque pulsation. In severe cases, they can even damage the motor itself and related equipment, causing economic losses and safety hazards. Therefore, developing diagnostic methods that can accurately and quickly identify the phase and type of open-circuit faults in multiphase motors has become a key requirement for ensuring the safe and stable operation of multiphase motors. The diagnostic approach of combined harmonic spatial current trajectory is based on the correlation between harmonic currents and fault states during multiphase motor operation. Through in-depth processing and analysis of harmonic currents, it explores technologies for fault diagnosis.

[0003] Existing technical solutions for open-circuit diagnosis of multiphase motors have two significant drawbacks. Firstly, some diagnostic methods rely solely on the fundamental current signal for fault diagnosis, failing to fully utilize the fault characteristic information contained in harmonic currents. This makes it difficult to capture effective fault signals in scenarios where the motor is operating under light load or in the early stages of a fault when harmonic components are weak, resulting in insufficient diagnostic accuracy as normal operation cannot accurately distinguish between normal operating conditions and minor open-circuit faults. Secondly, most diagnostic methods have poor adaptability to motor operating conditions. When multiphase motors are under complex conditions such as variable speed operation and load fluctuations, the current signal exhibits significant fluctuations. Existing methods lack effective signal processing and interference avoidance mechanisms, easily misjudging current fluctuations caused by changes in operating conditions as fault signals, or causing missed diagnoses due to fluctuations masking fault characteristics. Furthermore, they are difficult to adapt to multiphase motors with different numbers of phases and different topologies, limiting their applicability. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a multi-phase motor open circuit diagnosis method with combined harmonic spatial current trajectory.

[0005] The technical solution adopted in this invention is a multiphase motor open-circuit diagnosis method using combined harmonic spatial current trajectories, comprising: Step S1, filtering the multiphase winding currents collected during the operation of the multiphase motor, and obtaining two orthogonal current components corresponding to the 3rd harmonic subspace and two orthogonal current components corresponding to the 5th harmonic subspace through spatial coordinate transformation; Step S2, performing linear transformation on the two orthogonal current components of the 3rd harmonic subspace and the two orthogonal current components of the 5th harmonic subspace respectively to generate two sets of current variables corresponding to the number of motor phases, with the number of current variables in each set being consistent with the number of motor phases; Step S3, performing value inversion operations on the two sets of current variables respectively. Then, a moving average calculation is performed on the inverted current variable to obtain two sets of moving average results; Step S4: Set a threshold to avoid the influence of single-point trajectory on the judgment. If the trajectory corresponding to the moving average result meets the single-point characteristic, adjust the ratio of the two sets of moving average results so that the ratio deviates from the preset straight trajectory ratio range; Step S5: Define a judgment factor corresponding to the number of motor phases. Each judgment factor is calculated by the ratio of the two sets of moving average results of the corresponding phase, resulting in the same number of judgment factors as the number of motor phases; Step S6: Match the calculated judgment factors with the preset open circuit fault diagnosis table, and determine the open circuit fault phase and fault type of the multi-phase motor based on the matching result.

[0006] Furthermore, the spatial coordinate transformation process in step S1 satisfies the following relationship:

[0007]

[0008] The 6s / 2s transformation matrix in the formula is:

[0009] =

[0010] in, These represent the currents of phases a, b, c, d, e, and f, respectively. These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the fundamental frequency space; These represent the current components in two orthogonal directions in the third harmonic subspace; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

[0011] Furthermore, the calculation of the judgment factor in step S5 satisfies the following formula: ,in, The decision factor representing the nth phase, where n corresponds to... Mutually; This represents the moving average result of the nth phase in the 5th harmonic subspace; This represents the moving average result of the nth phase in the 3rd harmonic subspace; when A value of 0 results in When infinity occurs, The value is assigned to 3.

[0012] Furthermore, the preset range for matching the judgment factor in step S6 is determined in the following way: ,in, This represents the threshold used to adjust the judgment accuracy. This threshold is set according to the actual operating conditions of the multiphase motor and the current detection accuracy requirements. To determine the allowable fluctuation range of the standard value corresponding to the factor of 1; To determine the allowable fluctuation range of the standard value corresponding to the factor of 3; This is to determine the allowable fluctuation range of the standard value of -0.5 corresponding to the judgment factor.

[0013] Furthermore, the linear transformation process in step S2 satisfies the following relationship:

[0014]

[0015]

[0016] in, Representing the corresponding values ​​in the 3rd harmonic subspace The current component of the phase; These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the third harmonic subspace; Representing the corresponding values ​​in the 5th harmonic subspace The current component of the phase; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

[0017] Furthermore, the moving average calculation in step S3 satisfies the following formula:

[0018]

[0019]

[0020] in, The moving average result represents the inverted subspace current variable of the nth phase 3rd harmonic. The moving average result represents the inverted subspace current variable of the nth phase 5th harmonic; The window length representing the moving average calculation is set according to the operating frequency and current fluctuation of the multiphase motor. The value of the subspace current variable representing the 3rd harmonic of the nth phase at the sampling time; This represents the value of the 5th harmonic subspace current variable of the nth phase at the sampling time.

[0021] Further, step S3 includes the following sub-steps: S31 For the two sets of current variables generated in step S2, perform a judgment and sign reversal operation on the value of each current variable one by one, that is, if the value of the current variable is negative, it is converted to positive; if the value of the current variable is positive, it remains unchanged, completing the inversion process of all current variables; S32 Determine the window length for the moving average calculation. The window length is selected according to the sampling frequency of the multiphase motor and the dynamic change characteristics of the current signal to ensure that the fluctuation of the current signal can be effectively smoothed and the fault characteristic information is not lost; S33 Sequentially extract continuous sampling data of each set of inverted current variables according to the set window length, and perform arithmetic average calculation on the sampling data in each window; S34 Move the window sequentially according to the sampling time order and repeat the arithmetic average calculation process to obtain two sets of moving average results with the same number of sampling points as the original current variables. Each set of moving average results corresponds to the moving average data of the harmonic current of one phase.

[0022] Further, step S4 includes the following sub-steps: S41 Perform a one-to-one correspondence analysis on the two sets of moving average results obtained in step S3, and use the two sets of moving average results of the same phase as the horizontal and vertical coordinates of the coordinate points to construct the combined harmonic spatial current trajectory of each phase of the multiphase motor; S42 Observe the shape of the constructed combined harmonic spatial current trajectory. If the trajectory of a certain phase is a fixed single point, that is, the two sets of moving average results of that phase maintain a constant value and do not change at continuous sampling time; S43 Set a threshold to avoid the influence of single-point trajectory. The threshold is determined according to the fluctuation range of harmonic current during normal operation of the multiphase motor and the change amplitude of harmonic current during fault; S44 When a single point is detected in the trajectory of a certain phase, adjust the ratio of the two sets of moving average results of that phase according to the set threshold, so that the adjusted ratio is no longer within the ratio range corresponding to the preset 0°, 45°, and 63.43° straight line trajectories, so as to avoid the fault judgment deviation caused by the single-point trajectory.

[0023] Further, step S5 includes the following sub-steps: S51 Define judgment factors with the same number of phases as the multiphase motor, each judgment factor corresponding to one phase winding of the motor, clarifying the relationship between each judgment factor and the corresponding phase winding; S52 Obtain the two sets of moving average results obtained in step S3, and determine the two sets of moving average data corresponding to each judgment factor, namely the 5th harmonic subspace moving average result and the 3rd harmonic subspace moving average result of the phase corresponding to a certain judgment factor; S53 Calculate the value of each judgment factor by division, using the 5th harmonic subspace moving average result of the corresponding phase as the dividend and the 3rd harmonic subspace moving average result of the corresponding phase as the divisor, and calculate the ratio; S54 Check the calculated judgment factor values. If the judgment factor becomes infinitely large due to the divisor being 0, assign a value to the judgment factor according to the preset rules to ensure that each judgment factor has a valid value for fault diagnosis.

[0024] Furthermore, the multiphase motor is a six-phase open-winding permanent magnet synchronous motor, adopting a six-phase H-bridge topology with a common DC bus. Each phase winding is powered by a set of H-bridges including 4 IGBTs. The multiphase winding current collected in step S1 is the six-phase winding current. The two sets of current variables generated in step S2 each have 6 variables. The number of judgment factors defined in step S5 is 6. The preset open-circuit fault diagnosis table in step S6 includes judgment factor features corresponding to a total of 21 types of open-circuit faults, including single-phase open circuit, adjacent two-phase open circuit, one-phase-separated open circuit, and two-phase-separated open circuit.

[0025] Beneficial Effects: This invention proposes a multiphase motor open-circuit diagnostic method using a combined harmonic spatial current trajectory. By extracting the orthogonal current components of the 3rd and 5th harmonic subspaces and combining linear transformation, inversion, and moving average processing, it fully mines the fault characteristic information contained in the harmonic current, avoiding the problem of insufficient fault signal capture caused by relying solely on the fundamental current, and significantly improving diagnostic accuracy. Even in scenarios where the motor is operating under light load or the harmonic components are weak in the early stages of a fault, it can accurately identify fault characteristics by constructing a combined harmonic spatial current trajectory and defining judgment factors, thus solving the shortcomings of insufficient diagnostic accuracy in existing technologies. This method sets a threshold to avoid the influence of single-point trajectories, adjusts the matching range of judgment factors according to the motor operating conditions and detection accuracy requirements, and adapts to specific topologies such as six-phase open-winding permanent magnet synchronous motors. It includes the diagnosis of 21 types of open-circuit faults, effectively dealing with current signal fluctuations under complex operating conditions such as variable speed operation and load fluctuations, avoiding misjudgments and omissions. Furthermore, it can flexibly adjust the current variables and the number of judgment factors according to the number of motor phases, significantly improving adaptability to multiphase motors with different operating conditions and topologies. Attached Figure Description

[0026] Figure 1 This is a flowchart of the method steps of the present invention;

[0027] Figure 2 This is a diagram showing the combined harmonic current trajectory during an open-circuit fault according to the present invention. Detailed Implementation

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0029] like Figure 1 As shown, the multiphase motor open-circuit diagnostic method with combined harmonic space current trajectories includes:

[0030] Step S1: Filter the multiphase winding current collected during the operation of the multiphase motor, and obtain two orthogonal current components corresponding to the 3rd harmonic subspace and two orthogonal current components corresponding to the 5th harmonic subspace through spatial coordinate transformation.

[0031] Specifically, step S1 extracts harmonic current information reflecting open-circuit fault characteristics from the raw winding current collected during multiphase motor operation, providing a valid data foundation for subsequent fault diagnosis. In addition to the fundamental component, the winding current of a multiphase motor under normal operation and open-circuit fault conditions also generates harmonic components of different orders. Among them, the 3rd and 5th harmonics are highly sensitive to open-circuit faults and can clearly reflect the changes in current distribution caused by the fault; therefore, these two harmonic subspaces are selected for analysis. Through filtering, high-frequency noise, grid interference, and other irrelevant harmonic components mixed in with the raw current can be removed, avoiding these interference signals from affecting the accuracy of subsequent harmonic current component extraction. Spatial coordinate transformation utilizes the spatial distribution characteristics of the multiphase motor windings to transform the filtered multiphase current into a specific harmonic subspace, separating two orthogonal current components corresponding to the 3rd and 5th harmonic subspaces. These two orthogonal components can completely describe the magnitude and phase information of the current within the corresponding harmonic subspace, laying a data foundation for subsequently constructing a combined harmonic space current trajectory. In the specific implementation process, the winding current of the multiphase motor is first collected. Taking a six-phase open-winding permanent magnet synchronous motor as an example, the object of collection is the real-time current of its six-phase windings. The sampling frequency is set to 10kHz to 20kHz to ensure complete capture of the dynamic changes of the current signal. The sampling duration is determined according to the motor's operating cycle, usually 3 to 5 motor electrical cycles, to obtain sufficient current data for analysis. Subsequently, a filtering operation is performed using an infinite impulse response low-pass filter. The cutoff frequency of the filter is set to 500Hz to 800Hz. This frequency range can effectively retain the 3rd harmonic (assuming the motor's rated frequency is 50Hz, the 3rd harmonic frequency is 150Hz) and the 5th harmonic (250Hz) components, while filtering out high-frequency noise and interference signals with frequencies higher than 500Hz. A sliding window filtering method is used during the filtering process, with the window length set to 16 to 32 sampling points to further improve the smoothness of the filtered current signal. After filtering, a spatial coordinate transformation is performed. For a six-phase motor, the transformation matrix is ​​constructed based on the electrical angle distribution of each phase winding in space. The electrical angle interval between each phase winding is 30°. The transformation matrix converts the six-phase filtered current into two orthogonal current components in the 3rd harmonic subspace and two orthogonal current components in the 5th harmonic subspace. The numerical range of each orthogonal component is determined according to the rated current of the motor. For example, when the rated phase current of the motor is 10A, the numerical range of the orthogonal current component is usually between -15A and 15A, ensuring that the range of harmonic current variation under fault conditions can be covered.

[0032] Step S2: Perform linear transformations on the two orthogonal current components of the obtained 3rd harmonic subspace and the two orthogonal current components of the 5th harmonic subspace respectively to generate two sets of current variables corresponding to the number of motor phases. The number of current variables in each set is consistent with the number of motor phases.

[0033] Specifically, step S2 converts the orthogonal current components of the harmonic subspace extracted in step S1 into current variables directly corresponding to the number of motor phases. This allows subsequent fault characteristic analysis to be associated with specific phases of the motor, providing data support for accurate fault phase location. Since the orthogonal current components of the 3rd and 5th harmonic subspaces obtained in step S1 are abstract current information based on the overall space and cannot be directly mapped to each winding of the motor, and open-circuit faults typically occur in one or several specific phases, a linear transformation is needed to convert the abstract orthogonal components into current variables consistent with the number of motor phases. This ensures that each current variable corresponds to a single phase winding of the motor, enabling subsequent fault analysis targeting each phase winding. Simultaneously, the linear transformation can appropriately adjust the amplitude of the harmonic current components, ensuring that the current variables of different harmonic subspaces remain compatible in numerical range. This avoids deviations in subsequent calculations and judgments due to excessive differences in the amplitudes of different harmonic components, guaranteeing the quality of the constructed combined harmonic space current trajectory. In practice, the number of phases of the motor is first determined. Taking a six-phase motor as an example, the two orthogonal current components of the third harmonic subspace need to be converted into six current variables, and the two orthogonal current components of the fifth harmonic subspace also need to be converted into six current variables. Each set of current variables corresponds to phases a, b, c, d, e, and f of the six-phase motor.

[0034] Step S3: Invert the values ​​of the two sets of current variables respectively, and then perform a moving average calculation on the inverted current variables to obtain two sets of moving average results.

[0035] Specifically, step S3 preprocesses the current variables generated in step S2. Through inversion and moving average calculation, it eliminates local fluctuations and random interference in the current variables, highlighting the current change trend caused by the fault, and providing smooth data for constructing a stable and reliable combined harmonic space current trajectory. The inversion operation mainly adjusts the sign characteristics of the current variables to prevent mutual cancellation during subsequent moving average processing. The moving average calculation smooths out local numerical fluctuations in the current variables caused by minor fluctuations in motor load and detection noise during sampling. By averaging the current variables across multiple consecutive sampling times, it preserves the overall trend of the current variables while suppressing local fluctuations, enabling the obtained moving average result to more accurately reflect the actual operating state and fault characteristics of the motor, avoiding misjudgments in subsequent trajectory analysis due to local fluctuations. In practice, the inversion operation is first performed on the two sets of current variables generated in step S2 (six current variables corresponding to the 3rd harmonic subspace and six current variables corresponding to the 5th harmonic subspace), and the sign of each current variable is inverted one by one. For example, if the value of the current variable in phase a of the 3rd harmonic subspace is 5.2A, it will be inverted to -5.2A; if the value of the current variable in phase b of the 5th harmonic subspace is -3.8A, it will be inverted to 3.8A; if the value of a certain phase current variable is 0A, it will remain 0A after inversion. The inversion operation needs to traverse all current variables to ensure no omissions or errors. After inversion, a moving average calculation is performed. The window length of the moving average is set to 20 to 40 sampling points. The choice of window length is based on the sampling frequency of the motor and the current fluctuation frequency. For example, when the sampling frequency is 15kHz, the window length is set to 30 sampling points, and the corresponding time window is about 2ms, which can effectively smooth fluctuations and reflect the current change trend in a timely manner. During the moving average calculation, starting from the first sampling time, continuous inverted current variable data are extracted with a set window length. The arithmetic mean of the data within the window is calculated to obtain the first moving average result. Then, the window is moved forward by one sampling point, and the above averaging calculation process is repeated until all sampled data are traversed, finally obtaining two sets of moving average results. The number of values ​​included in each set is the same as the number of sampling points of the original current variable. The numerical precision of each moving average result is retained to two decimal places. For example, the moving average result of the third harmonic subspace a phase may be a continuous value such as 2.35A, 2.32A, 2.36A, etc.

[0036] Step S4: Set a threshold to avoid the influence of single-point trajectory on the judgment. If the trajectory corresponding to the moving average result meets the single-point feature, adjust the ratio of the two sets of moving average results so that the ratio is outside the preset range of straight trajectory ratio.

[0037] Specifically, step S4 addresses the potential single-point trajectory problem in the combined harmonic space current trajectory, preventing deviations in subsequent fault judgment due to single-point trajectories and ensuring the accuracy of diagnostic results. During multiphase motor operation, certain special operating conditions (such as extremely low motor speed, constant and unfluctuating load) or accidental anomalies during current detection may cause the two sets of moving average results for a certain phase to remain constant at continuous sampling times, forming a fixed single-point trajectory. This single-point trajectory cannot reflect the normal variation pattern and fault characteristics of the phase current. If directly used for subsequent judgment, it may misjudge normal operating conditions as faults or mask actual faults. Therefore, it is necessary to set a specific threshold to identify single-point trajectories and adjust them so that the adjusted trajectory can move away from the easily confused specific straight-line trajectory ratio range, restoring the normal fault characteristic recognition capability and ensuring the effectiveness of subsequent judgment factor calculation and fault matching. In specific implementation, the two sets of moving average results obtained in step S3 are first correlated. The 3rd harmonic subspace moving average result of the same phase is used as the abscissa, and the 5th harmonic subspace moving average result is used as the ordinate to construct the combined harmonic spatial current trajectory of that phase. For example, for a six-phase motor, six trajectories need to be constructed for phases a to f respectively. Then, a threshold for identifying single-point trajectories is set. This threshold includes a numerical fluctuation threshold and a duration threshold. The numerical fluctuation threshold is set to 0.05A to 0.1A, that is, if the numerical change of the two sets of moving average results of a certain phase is less than the set fluctuation threshold at consecutive sampling times; at the same time, the duration threshold is set to 5 to 10 sampling times, that is, if this stable numerical state continues to reach the set time threshold, then the trajectory of that phase is determined to be a single-point trajectory. When a single-point trajectory is detected, the ratio of the two sets of moving average results is adjusted. First, a preset range of ratios for specific straight-line trajectories is determined. This range is based on the distribution of the ratios of the two sets of moving average results when the motor is running normally. For example, the ratio range for a 0° straight-line trajectory is 0.95 to 1.05, the ratio range for a 45° straight-line trajectory is 1.95 to 2.05, and the ratio range for a 63.43° straight-line trajectory is 2.95 to 3.05. During adjustment, a small numerical adjustment is made to one of the two sets of moving average results according to the set threshold (usually 0.1 to 0.3). For example, if the moving average result of the 3rd harmonic subspace is 2.5A and the moving average result of the 5th harmonic subspace is 2.5A, and the ratio is 1.0 (within the 0° straight trajectory range), then the moving average result of the 5th harmonic subspace is adjusted to 2.7A, so that the adjusted ratio becomes 1.08, which is outside the range of 0.95 to 1.05. During the adjustment process, it is ensured that the numerical adjustment amount does not affect the overall current characteristics and is only used to avoid the influence of single-point trajectory.

[0038] Step S5: Define the judgment factors corresponding to the number of motor phases. Each judgment factor is calculated by the ratio of the two sets of moving average results of the corresponding phase, so that the number of judgment factors is the same as the number of motor phases.

[0039] Specifically, step S5 transforms the moving average result obtained in step S3 into a judgment factor that can be directly used for fault diagnosis, establishing a correlation between the moving average data and fault characteristics, and providing quantitative indicators for subsequent fault matching. Since the two sets of moving average results obtained in step S3 correspond to current information in different harmonic subspaces, analyzing only one set of results is insufficient to directly determine the fault condition. However, by calculating the ratio of the two sets of results, the current characteristics of the two harmonic subspaces can be integrated to form a judgment factor with clear fault differentiation. Under normal operation and open-circuit fault conditions, the ratio of the two sets of moving average results for different phase windings will exhibit different numerical characteristics. For example, the ratio of the normal phase may be stable within a certain fixed range, while the ratio of the faulty phase will deviate from this range. Therefore, a number of judgment factors equal to the number of motor phases are defined, with each judgment factor corresponding to one phase winding. This enables individual judgment of the fault state of each phase winding, providing a crucial basis for accurately locating the faulty phase. In specific implementation, the number of judgment factors is first determined based on the number of motor phases. For example, a six-phase motor requires six judgment factors, labeled as the judgment factors corresponding to phases a to f. Subsequently, the two sets of moving average results obtained in step S3 are acquired, clarifying the two sets of data corresponding to each judgment factor, namely, the 5th harmonic subspace moving average result (as the numerator) and the 3rd harmonic subspace moving average result (as the denominator) of the phase corresponding to a certain judgment factor. During the calculation process, the ratio of each judgment factor is calculated one by one. For example, when calculating the judgment factor of phase a, the 5th harmonic subspace moving average result of phase a (e.g., 2.8A) is divided by the 3rd harmonic subspace moving average result of phase a (e.g., 1.4A), resulting in a judgment factor value of 2.0 for phase a; when calculating the judgment factor of phase b, if the 5th harmonic subspace moving average result of phase b is 4.5A and the 3rd harmonic subspace moving average result is 1.5A, then the judgment factor value of phase b is 3.0. During the calculation process, it is necessary to check in real time whether the denominator (the moving average result of the 3rd harmonic subspace) is 0. If the denominator is 0 (e.g., the moving average result of the 3rd harmonic subspace of a certain phase is 0A), the judgment factor of that phase is assigned a value of 3 according to the preset rules (this value is determined based on a large amount of fault experimental data and can effectively distinguish fault states), so as to avoid infinite values ​​affecting subsequent diagnosis. The calculation result of each judgment factor is retained to one decimal place to ensure that the numerical accuracy meets the fault judgment requirements. After the calculation is completed, all judgment factors are summarized to form a complete judgment factor set. For example, the judgment factor set of a six-phase motor may be [2.0, 3.0, 1.9, 2.1, 3.1, 2.0].

[0040] Step S6: Match the calculated judgment factor with the preset open circuit fault diagnosis table, and determine the open circuit fault phase and fault type of the multiphase motor based on the matching result.

[0041] Specifically, step S6 is the final step in the entire diagnostic method. By matching the judgment factors calculated in step S5 with the preset open-circuit fault diagnosis table, the phase and specific fault type of the open-circuit fault in the multi-phase motor are determined, achieving accurate fault location and identification. The preset open-circuit fault diagnosis table is constructed based on a large amount of experimental data on open-circuit faults in multi-phase motors, including various possible open-circuit fault situations (such as single-phase open circuit, two-phase open circuit, etc.), and the typical numerical range or characteristic patterns of the judgment factors for each phase under each fault condition. Through the matching operation, the current motor's judgment factor characteristics can be compared with the fault characteristics in the diagnosis table to find the most consistent fault type, thereby quickly and accurately determining whether the motor has an open-circuit fault and which phase windings the fault occurs in. This provides a clear basis for subsequent fault repair and motor protection, avoiding repair delays or incorrect repairs caused by inaccurate fault location. In practice, the pre-set open-circuit fault diagnosis table is first retrieved. This table is designed for specific types of multiphase motors (such as six-phase open-winding permanent magnet synchronous motors) and includes 21 common open-circuit fault types, including 6 single-phase open-circuit faults (phases a to f are open-circuited separately) and 15 two-phase open-circuit faults (two adjacent phases, one phase apart, and two phases apart). The diagnosis table records the normal value range of the judgment factor for each phase under each fault type. For example, during normal operation, the judgment factor range for each phase is 1.9 to 2.1. When phase a is open-circuited, the judgment factor range for phase a is 2.9 to 3.1, the judgment factor range for phase b is 1.4 to 1.6, and the other phases remain at 1.9 to 2.1. When two adjacent phases (phases a and b) are open-circuited, the judgment factor range for phase a is 2.9 to 3.1, the judgment factor range for phase b is 2.9 to 3.1, and the other phases range are 1.4 to 1.6, etc. Subsequently, the set of judgment factors calculated in step S5 is matched one by one with the range of judgment factors corresponding to each fault type in the diagnostic table. The matching degree between the current set of judgment factors and the range of judgment factors for each fault type is calculated. The matching degree is based on the proportion of judgment factors that match the range. For example, if 5 out of 6 judgment factors match the range of a certain fault type, the matching degree is 83.3%. A matching degree threshold of 80% to 90% is set. When the matching degree of a certain fault type reaches or exceeds the set threshold, it is determined that the motor currently has an open circuit fault of that type, and the corresponding fault phase is determined. If the matching degree of all fault types is lower than the threshold, it is determined that the motor currently has no open circuit fault, or there is a special fault situation not included in the diagnostic table, which requires further investigation. After the matching is completed, the fault diagnosis result is output, clarifying the fault phase (e.g., phase a) and fault type (e.g., single-phase open circuit), providing a basis for subsequent processing.

[0042] Preferably, the spatial coordinate transformation process in step S1 satisfies the following relationship:

[0043]

[0044] The 6s / 2s transformation matrix in the formula is:

[0045] =

[0046] in, These represent the currents of phases a, b, c, d, e, and f, respectively. These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the fundamental frequency space; These represent the current components in two orthogonal directions in the third harmonic subspace; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

[0047] Specifically, the spatial coordinate transformation process in step S1 clarifies the specific relationships followed by the transformation. Through precise mathematical relationships of coordinate transformation, it ensures that the orthogonal current components of the 3rd and 5th harmonic subspaces separated from the multiphase winding current accurately reflect the actual harmonic current state of each phase winding, laying a precise data foundation for subsequent fault feature extraction. In specific implementation, it is first necessary to clarify the definitions of various parameters involved in the transformation process. Among them, the 3rd harmonic subspace current components corresponding to different phases are the current manifestations of each phase winding in the 3rd harmonic subspace, and their magnitude is directly related to the degree of current distortion of the phase winding under fault conditions. The initial phase angle of each phase current is determined by the winding structure and spatial distribution of the multiphase motor. For example, the electrical angle interval between each phase winding of a six-phase motor is usually 30°, so the initial phase angles of each phase differ by 30°. The accuracy of this parameter directly affects the accuracy of the coordinate transformation results. The orthogonal current components of the 3rd and 5th harmonic subspaces are the current components in mutually perpendicular directions within the corresponding harmonic subspaces, which can completely describe the vector characteristics of the current within the harmonic subspace. In practical applications, parameter determination requires consideration of the specific motor model and design parameters, such as the motor's rated power, rated speed, and number of winding turns. Accurate values ​​of the initial phase angle of each phase are obtained through electromagnetic simulation or experimental measurement. The numerical range of each phase harmonic current component needs to be determined based on the current variation patterns under normal motor operation and fault conditions, typically between -1.5 and 1.5 times the rated phase current of the motor, to ensure coverage of the current variation range during faults. Through this spatial coordinate transformation, the complex current signal of multi-phase windings can be decomposed into clear orthogonal components in the harmonic subspace, avoiding mutual interference between different harmonic components. This makes subsequent analysis of each harmonic subspace more targeted, effectively improving the accuracy of fault feature extraction and providing crucial support for the reliability of the entire diagnostic method.

[0048] Preferably, the calculation of the judgment factor in step S5 satisfies the following formula:

[0049] in, The decision factor representing the nth phase, where n corresponds to... Mutually; This represents the moving average result of the nth phase in the 5th harmonic subspace; This represents the moving average result of the nth phase in the 3rd harmonic subspace; when A value of 0 results in When infinity occurs, The value is assigned to 3.

[0050] Specifically, in step S5, the calculation of the judgment factor involves defining its specific formula. The core function of this formula is to transform two independent sets of current data into a quantitative indicator with clear fault discrimination by fusing the moving average results of the 3rd and 5th harmonic subspaces, thus achieving accurate characterization of the fault state of each phase winding. During implementation, it is first necessary to clearly define each parameter in the formula. The judgment factor for the nth phase is the core indicator used to determine whether an open-circuit fault exists in that phase, and its value directly reflects the degree of fault in that phase. The moving average result of the nth phase in the 5th harmonic subspace is the result of inverting and averaging the 5th harmonic current of that phase, which can smoothly reflect the overall trend of the 5th harmonic current of that phase. Under fault conditions, this value will show significant fluctuations deviating from the normal range. Similarly, the moving average result of the nth phase in the 3rd harmonic subspace reflects the smooth trend of the 3rd harmonic current of that phase. When setting parameters, a reasonable numerical range needs to be determined based on a large amount of experimental data. For example, under normal operating conditions, the judgment factor is usually stable between 1.9 and 2.1. When an open-circuit fault occurs in a phase, the judgment factor for that phase will deviate from this range; for example, it may rise to between 2.9 and 3.1 when a single phase is open-circuited. Simultaneously, special handling is required for the special case where the third harmonic subspace moving average result is 0. In this case, directly calculating the ratio would result in infinity, rendering the judgment factor unusable. Therefore, the judgment factor in this case is preset to 3. This value is calibrated based on a large amount of fault experimental data, avoiding interference from infinite values ​​in subsequent diagnosis while conforming to the numerical characteristics of the judgment factor under fault conditions, ensuring the continuity and accuracy of the diagnostic process. In practical applications, the rationality of this assignment needs to be verified through multiple experiments. The value should be fine-tuned based on the experimental results of different motor models. For example, for multi-phase motors with lower power, the assignment can be adjusted to between 2.8 and 3.2 to adapt to their current variation characteristics. The judgment factor obtained through this calculation method can effectively distinguish between normal and faulty phases, providing a clear and reliable quantitative basis for fault matching in step S6.

[0051] Preferably, the preset range matching the judgment factor in step S6 is determined in the following way: ,in, This represents the threshold used to adjust the judgment accuracy. This threshold is set according to the actual operating conditions of the multiphase motor and the current detection accuracy requirements. To determine the allowable fluctuation range of the standard value corresponding to the factor of 1; To determine the allowable fluctuation range of the standard value corresponding to the factor of 3; This is to determine the allowable fluctuation range of the standard value of -0.5 corresponding to the judgment factor.

[0052] Specifically, in step S6, the preset range for the judgment factor matching is determined by clarifying its method. By setting a reasonable allowable fluctuation range, misjudgments caused by minor fluctuations in the current signal are avoided while ensuring diagnostic accuracy. Simultaneously, it ensures accurate capture of changes in the judgment factor during faults, achieving reliable fault identification. During implementation, the definitions and functions of each parameter must first be clarified. The threshold used to adjust the judgment accuracy is a key parameter determining the width of the preset range; its value must be determined comprehensively based on the actual operating conditions of the multiphase motor and the accuracy of current detection. The three different preset ranges correspond to typical values ​​of the judgment factor under normal operation and different fault conditions. The allowable fluctuation range corresponding to 1 is the reasonable range for the judgment factor under normal operation; the range corresponding to 3 is the common range for the judgment factor when an open-circuit fault occurs in a phase; and the range corresponding to -0.5 may correspond to the judgment factor performance under specific types of faults or special operating conditions. Regarding parameter settings, the threshold used to adjust the judgment accuracy is typically set between 0.05 and 0.1. For example, when the motor operating environment is relatively stable and the current detection accuracy is high, the threshold can be set to 0.05 to narrow the preset range and improve diagnostic accuracy. When the motor operating environment is complex and the current fluctuates significantly, the threshold can be adjusted to 0.1 to expand the allowable fluctuation range and avoid misjudgments. The center values ​​of each preset range (1, 3, -0.5) need to be obtained through statistical analysis of a large amount of experimental data. For example, normal operation and fault simulation experiments are conducted on multiple motors of the same model, and the average value of the judgment factor under each state is recorded. This average value is used as the center value of the preset range. The width of the range is determined by the threshold for adjusting the accuracy. The specific range interval is obtained by adding or subtracting the threshold from the center value. In practical applications, the preset range also needs to be dynamically adjusted according to the motor's operating time and aging degree. For example, after the motor has been running for many years, the winding parameters may change, and the judgment factor during normal operation may have a slight deviation. At this time, it is necessary to re-determine the new center value and fluctuation range through experimental measurement to ensure that the preset range always matches the actual operating state of the motor, providing an accurate judgment standard for fault matching in step S6, and effectively balancing diagnostic accuracy and anti-interference capability.

[0053] Preferably, the linear transformation process in step S2 satisfies the following relationship:

[0054]

[0055]

[0056] in, Representing the corresponding values ​​in the 3rd harmonic subspace The current component of the phase; These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the third harmonic subspace; Representing the corresponding values ​​in the 5th harmonic subspace The current component of the phase; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

[0057] Specifically, the linear transformation process in step S2 involves establishing a clear linear transformation mathematical model to accurately convert the orthogonal current components of the harmonic subspace into current variables corresponding to the number of motor phases. This ensures that the converted current variables meet subsequent processing requirements and improves the identifiability of fault characteristics. During implementation, detailed definitions of various parameters are necessary. The results of the linear transformation of the 3rd and 5th harmonic subspace current components are directly used for subsequent inversion and moving average operations, and their values ​​must be controlled within a range conducive to processing. In practice, parameter determination must be based on the specific motor model and design parameters, including the motor's rated power, rated speed, and number of winding turns. Accurate values ​​of the initial phase angle of each phase are obtained through electromagnetic simulation or experimental measurement. The numerical range of each phase harmonic current component should be determined based on the current variation patterns of the motor under normal operation and fault conditions. This range is typically set to -1.5 to 1.5 times the rated phase current to cover potential current fluctuations under fault conditions. This linear transformation relationship can convert abstract orthogonal current components into numerically compatible current variables associated with specific phases, providing high-quality data for the smooth implementation of subsequent steps and effectively improving the efficiency and accuracy of the entire diagnostic process.

[0058] Preferably, the moving average calculation in step S3 satisfies the following formula:

[0059]

[0060]

[0061] in, The moving average result represents the inverted subspace current variable of the nth phase 3rd harmonic. The moving average result represents the inverted subspace current variable of the nth phase 5th harmonic; The window length representing the moving average calculation is set according to the operating frequency and current fluctuation of the multiphase motor. The value of the subspace current variable representing the 3rd harmonic of the nth phase at the sampling time; This represents the value of the 5th harmonic subspace current variable of the nth phase at the sampling time.

[0062] Specifically, the formula for calculating the moving average in step S3 uses precise mathematical calculations to smooth the inverted current variable, eliminate local fluctuations and noise interference, and retain the overall fault characteristic trend of the current variable, providing reliable data for the subsequent construction of a stable combined harmonic spatial current trajectory. During implementation, it is necessary to clearly define each parameter in the formula. The moving average results of the 3rd and 5th harmonic subspaces of the nth phase are the final data of the corresponding harmonic current after inversion and moving average. They can accurately reflect the long-term variation law of the harmonic current of the phase and will show numerical characteristics that are significantly different from those of the normal state under fault conditions. The window length of the moving average calculation is a key parameter that determines the smoothing effect. Its value needs to be determined comprehensively based on the sampling frequency of the motor and the current fluctuation frequency. It is necessary to ensure that high-frequency fluctuations can be effectively smoothed while avoiding over-smoothing that would lead to the loss of fault characteristics. The value of the current variable of the corresponding harmonic subspace of the nth phase at the kth sampling time is the original data for the moving average calculation. Its accuracy directly affects the accuracy of the average result. The negative sign corresponds to the inversion operation in step S3 and is used to adjust the sign characteristics of the current variable to enhance the fault discrimination of the subsequent ratio calculation. Regarding parameter settings, the window length is typically set to 20 to 40 sampling points. For example, when the motor sampling frequency is 15kHz and the current fluctuation frequency is 500Hz, a window length of 30 sampling points corresponds to a time window of approximately 2ms. This window length effectively filters high-frequency fluctuations above 500Hz while fully preserving the low-frequency current change trend caused by faults. The values ​​of the nth phase current variable at each sampling time need to be acquired using a high-precision current sensor, with sampling accuracy typically controlled within ±0.01A to ensure the accuracy of the original data. In the actual calculation process, data within the window needs to be extracted sequentially according to the sampling time order. The arithmetic mean of the inverted current variable within each window is then calculated. For example, for a window with 30 sampling points, the 30 inverted current values ​​are added together and then divided by 30 to obtain the moving average result for that window. The window is then moved forward by one sampling point, and the above calculation is repeated until all sampling data has been traversed. This moving average calculation method can significantly reduce noise interference in the current signal, making the obtained moving average result more reflective of the motor's true operating status and fault characteristics. It avoids deviations in subsequent trajectory analysis and judgment factor calculation due to local fluctuations, thus providing an important guarantee for the reliability of the entire diagnostic method.

[0063] Preferably, step S3 includes the following sub-steps: S31 For the two sets of current variables generated in step S2, perform a judgment and sign reversal operation on the value of each current variable one by one, that is, if the value of the current variable is negative, it is converted to positive; if the value of the current variable is positive, it remains unchanged, and the inversion processing of all current variables is completed; S32 Determine the window length for the moving average calculation. The window length is selected according to the sampling frequency of the multiphase motor and the dynamic change characteristics of the current signal to ensure that the fluctuation of the current signal can be effectively smoothed and the fault characteristic information is not lost; S33 Sequentially extract continuous sampling data of each set of inverted current variables according to the set window length, and perform arithmetic average calculation on the sampling data in each window; S34 Move the window sequentially according to the sampling time order and repeat the arithmetic average calculation process to obtain two sets of moving average results with the same number of sampling points as the original current variables. Each set of moving average results corresponds to the moving average data of the harmonic current of one phase.

[0064] Specifically, step S3 includes four sub-steps. Through a structured operation process, it ensures that the inversion of current variables and the calculation of moving averages are executed accurately and systematically. This eliminates local fluctuations in the current signal while fully preserving fault characteristics, providing high-quality data for subsequent trajectory construction. During implementation, the current variable inversion operation is performed first. Each value in the two sets of current variables generated in step S2 (corresponding to the 3rd and 5th harmonic subspaces, respectively) is judged and its sign reversed to ensure no omissions or errors. This operation adjusts the sign characteristics of the current variables, preventing mutual cancellation during subsequent moving average processing. Next, the moving average window length is determined. This parameter needs to be selected based on the motor sampling frequency and current dynamic characteristics, typically between 20 and 40 sampling points. For example, 30 sampling points are selected at a sampling frequency of 15kHz to smooth fluctuations without losing fault information. Then, the inverted current data is truncated according to the window length and an arithmetic average is calculated to ensure accurate calculation within each window. Finally, the window is moved chronologically and the calculation is repeated to obtain two sets of moving average results with the same number of sampling points as the original, each set corresponding to one phase harmonic current data. The entire process must strictly follow the step-by-step logic. When inverting, the accuracy of the sign conversion must be verified. The window length selection must be verified for adaptability through experiments. When calculating the average, the numerical precision must be controlled (two decimal places must be retained). Through this refined process, data errors caused by disordered operation can be avoided, ensuring that the moving average result can accurately reflect the motor's operating status and fault characteristics, and laying a reliable data foundation for subsequent steps.

[0065] Preferably, step S4 includes the following sub-steps: S41 Perform a one-to-one correspondence analysis on the two sets of moving average results obtained in step S3, and use the two sets of moving average results of the same phase as the horizontal and vertical coordinates of the coordinate points to construct the combined harmonic spatial current trajectory of each phase of the multiphase motor; S42 Observe the shape of the constructed combined harmonic spatial current trajectory. If the trajectory of a certain phase is a fixed single point, that is, the two sets of moving average results of that phase maintain a constant value and do not change at continuous sampling time; S43 Set a threshold to avoid the influence of single-point trajectory. The threshold is determined according to the fluctuation range of harmonic current during normal operation of the multiphase motor and the change amplitude of harmonic current during fault; S44 When a single point is detected in the trajectory of a certain phase, adjust the ratio of the two sets of moving average results of that phase according to the set threshold so that the adjusted ratio is no longer within the ratio range corresponding to the preset 0°, 45°, and 63.43° straight line trajectories, so as to avoid the fault judgment deviation caused by the single-point trajectory.

[0066] Specifically, step S4 is further refined into steps. Through a systematic trajectory analysis and adjustment process, the interference of single-point trajectories on diagnosis is accurately identified and avoided, ensuring that the combined harmonic space current trajectory can effectively reflect the fault characteristics and avoid misjudgment or omission. In implementation, firstly, the two sets of moving average results obtained in step S3 are used to construct the trajectory of each phase with "the third harmonic result of the same phase as the horizontal axis and the fifth harmonic result as the vertical axis", clearly presenting the spatial distribution characteristics of the current variable. Next, the trajectory shape is observed to determine whether there is a fixed single point (the value remains constant and unchanged during continuous sampling). This step requires phase-by-phase analysis to ensure that no single point trajectory is missed. Then, an avoidance threshold is set. This threshold needs to be determined by combining the range of harmonic current fluctuations during normal motor operation and the change amplitude during faults. Typically, the numerical fluctuation threshold is 0.05 to 0.1, and the duration threshold is 5 to 10 sampling times to ensure accurate identification of single point trajectories. Finally, when a single point trajectory is detected, the ratio of the two sets of moving average results of the corresponding phase is adjusted according to the threshold to make it deviate from the ratio range of the 0°, 45°, and 63.43° straight trajectories. During the adjustment process, the numerical change amplitude needs to be controlled to eliminate the influence of single points without changing the overall fault characteristics. During implementation, trajectory construction must ensure accurate correspondence between horizontal and vertical coordinates; single-point judgment must combine fluctuation and time thresholds to avoid misjudgment; threshold setting must be calibrated through a large number of experiments; and adjustment operations must verify whether the ratio deviates from the target range. Through this refined process, the diagnostic bias caused by single-point trajectories can be effectively resolved, ensuring the accuracy of subsequent judgment factor calculation and fault matching.

[0067] Preferably, step S5 includes the following sub-steps: S51 Define a judgment factor with the same number of phases as the multiphase motor, with each judgment factor corresponding to one phase winding of the motor, clarifying the relationship between each judgment factor and the corresponding phase winding; S52 Obtain the two sets of moving average results obtained in step S3, and determine the two sets of moving average data corresponding to each judgment factor, namely the 5th harmonic subspace moving average result and the 3rd harmonic subspace moving average result of the phase corresponding to a certain judgment factor; S53 Calculate the value of each judgment factor by division, using the 5th harmonic subspace moving average result of the corresponding phase as the dividend and the 3rd harmonic subspace moving average result of the corresponding phase as the divisor, and calculate the ratio; S54 Check the calculated judgment factor values. If the judgment factor becomes infinitely large due to the divisor being 0, assign a value to the judgment factor according to the preset rules to ensure that each judgment factor has a valid value for fault diagnosis.

[0068] Specifically, step S5, consisting of four sub-steps, transforms the moving average result into a quantitative indicator with clear fault differentiation through standardized judgment factor definition, data matching, calculation, and anomaly handling processes, providing a precise basis for fault diagnosis. During implementation, firstly, the number of judgment factors is defined according to the number of motor phases (e.g., 6 for a six-phase motor), and the correlation between each factor and its corresponding phase winding is clarified to ensure a one-to-one correspondence between factors and phases, avoiding confusion. Next, the two sets of moving average results from step S3 are obtained, determining the 5th and 3rd harmonic subspace data corresponding to each judgment factor. This step requires strict phase matching to ensure accurate data sources. Then, the ratio is calculated using the 5th harmonic data as the dividend and the 3rd harmonic data as the divisor to obtain the judgment factor value. Precision must be controlled during calculation (retaining one decimal place) to ensure the value reflects fault differences. Finally, the calculation result is checked. If the 3rd harmonic data is 0, resulting in infinity, a value is assigned according to a preset rule (usually 3, requiring experimental calibration) to avoid abnormal values ​​affecting diagnosis. During implementation, the number of judgment factors needs to be determined in conjunction with the motor structure, data matching needs to be checked phase by phase, calculation operations need to verify the division logic, and anomaly handling needs to verify the rationality of the assignment. Through this refined process, it can be ensured that the value of each judgment factor is accurate and effective, which not only avoids data association errors but also solves the interference of outliers, providing a reliable quantitative indicator for fault matching in step S6.

[0069] Preferred, such as Figure 2As shown, the multiphase motor is a six-phase open-winding permanent magnet synchronous motor, adopting a six-phase H-bridge topology with a common DC bus. Each phase winding is powered by an H-bridge consisting of four IGBTs. The multiphase winding current collected in step S1 is the six-phase winding current. The two sets of current variables generated in step S2 each contain six variables. The number of judgment factors defined in step S5 is six. The preset open-circuit fault diagnosis table in step S6 includes judgment factor features corresponding to 21 types of open-circuit faults, namely, single-phase open circuit, adjacent two-phase open circuit, one-phase-separated open circuit, and two-phase-separated open circuit. Figure 2 The seven cases are (a) to (g).

[0070] Specifically, the applicable motor is clearly defined as a six-phase open-winding permanent magnet synchronous motor, using a six-phase H-bridge topology with a common DC bus. Each phase is powered by an H-bridge containing four IGBTs. This limitation ensures that the method is adapted to a specific motor structure, guaranteeing that the technical solution matches the motor characteristics. Next, corresponding to the number of motor phases, steps S1 (collecting the six-phase winding current), S2 (generating six current variables per group), and S5 (defining six judgment factors) are specified to ensure that the parameters of each step are adapted to the number of motor phases, avoiding errors caused by parameter mismatch. Finally, the preset diagnostic table in step S6 includes 21 open-circuit fault types (6 single-phase and 15 two-phase), including adjacent, one-phase-separated, and two-phase-separated open circuits, ensuring that the method can diagnose multiple fault conditions and improve its practicality. During implementation, the limitations of motor type and topology need to be combined with common industry application scenarios, the parameters of each step need to be precisely set according to the number of motor phases, and the range of fault types needs to be sorted out through fault simulation experiments to ensure coverage of major fault conditions. Through these limitations, the method is transformed from a general framework into a practical solution for specific motors, which not only ensures diagnostic accuracy but also expands its application in the field of six-phase open-winding permanent magnet synchronous motors, meeting the fault diagnosis needs of specific motors in industrial scenarios.

[0071] A combined harmonic spatial current trajectory diagnostic method for multiphase motors offers the primary advantage of comprehensive fault feature capture, effectively addressing the shortcomings of existing technologies that rely solely on the fundamental current, resulting in insufficient diagnostic accuracy. This method extracts orthogonal current components from specific harmonic subspaces, performs linear transformations to generate current variables matching the number of motor phases, and then combines inversion and moving average calculations to fully uncover hidden fault information within the harmonic currents. Compared to traditional methods that only focus on the fundamental signal, this in-depth processing of harmonic signals preserves fault characteristics even under light motor loads or in the early stages of a fault when harmonic components are weak, preventing missed diagnoses due to insufficient signal capture and significantly improving diagnostic accuracy. Secondly, this method possesses excellent anti-interference capabilities, addressing misdiagnosis issues caused by current fluctuations under complex operating conditions in existing technologies. During processing, a specific threshold is set to avoid single-point trajectories. When a phase trajectory is detected to exhibit a fixed single-point pattern, the ratio of the two moving average results is adjusted to move it out of the range of easily confused straight-line trajectory ratios. Meanwhile, the matching range of the judgment factors can be flexibly adjusted according to the actual operating conditions of the motor and the requirements of detection accuracy. This effectively filters interference from current signals under conditions such as variable speed operation and load fluctuations, avoiding misjudging normal fluctuations as faults or judgment deviations caused by interference masking fault characteristics, thus ensuring the reliability of diagnostic results under complex operating conditions. Finally, this method has wide adaptability, overcoming the shortcomings of existing technologies in terms of poor adaptability to motors with different topologies and phase numbers. Its design allows adjustment of the number of current variables and judgment factors according to the number of motor phases. For example, for a six-phase open-winding permanent magnet synchronous motor, it can generate a corresponding number of current variables and judgment factors, including diagnoses of various open-circuit fault types. This flexible adaptability design makes it not only applicable to multi-phase motors with specific topologies, but also allows adjustment of diagnostic parameters and matching standards according to the structural characteristics and fault type requirements of different motors, eliminating the need to develop separate diagnostic schemes for each motor and significantly expanding the application scope of the method.

[0072] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for diagnosing open-circuit faults in multiphase motors using combined harmonic spatial current trajectories, characterized in that, include: Step S1: Filter the multiphase winding current collected during the operation of the multiphase motor, and obtain two orthogonal current components corresponding to the 3rd harmonic subspace and two orthogonal current components corresponding to the 5th harmonic subspace through spatial coordinate transformation. Step S2: Perform linear transformations on the two orthogonal current components of the obtained 3rd harmonic subspace and the two orthogonal current components of the 5th harmonic subspace respectively to generate two sets of current variables corresponding to the number of motor phases. The number of current variables in each set is consistent with the number of motor phases. Step S3: Invert the values ​​of the two sets of current variables, and then perform a moving average calculation on the inverted current variables to obtain two sets of moving average results; the moving average calculation satisfies the following formula: in, The moving average result represents the inverted subspace current variable of the nth phase 3rd harmonic. The moving average result represents the inverted subspace current variable of the nth phase 5th harmonic; The window length representing the moving average calculation is set according to the operating frequency and current fluctuation of the multiphase motor. The value of the subspace current variable representing the 3rd harmonic of the nth phase at the sampling time; The value of the subspace current variable representing the 5th harmonic of the nth phase at the sampling time; Step S4: Set a threshold to avoid the influence of single-point trajectory on the judgment. If the trajectory corresponding to the moving average result meets the single-point feature, adjust the ratio of the two sets of moving average results so that the ratio is outside the preset range of straight trajectory ratio. Step S5: Define the judgment factors corresponding to the number of motor phases. Each judgment factor is calculated by the ratio of the moving average results of two sets of corresponding phases, resulting in the same number of judgment factors as the number of motor phases. The calculation of the judgment factors satisfies the following formula: , in, The decision factor representing the nth phase, where n corresponds to... Mutually; This represents the moving average result of the nth phase in the 5th harmonic subspace; This represents the moving average result of the nth phase in the 3rd harmonic subspace; when A value of 0 results in When infinity occurs, The value is assigned to 3; Step S6: Match the calculated judgment factors with the preset open circuit fault diagnosis table, and determine the open circuit fault phase and fault type of the multiphase motor based on the matching results. The preset range for matching the judgment factor is determined in the following way: ,in, This represents the threshold used to adjust the judgment accuracy. This threshold is set according to the actual operating conditions of the multiphase motor and the current detection accuracy requirements. To determine the allowable fluctuation range of the standard value corresponding to the factor of 1; To determine the allowable fluctuation range of the standard value corresponding to the factor of 3; This is to determine the allowable fluctuation range of the standard value of -0.5 corresponding to the judgment factor.

2. The multiphase motor open-circuit diagnosis method with combined harmonic spatial current trajectory according to claim 1, characterized in that, The spatial coordinate transformation process in step S1 satisfies the following relationship: The 6s / 2s transformation matrix in the formula is: = in, These represent the currents of phases a, b, c, d, e, and f, respectively. These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the fundamental frequency space; These represent the current components in two orthogonal directions in the third harmonic subspace; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

3. The multiphase motor open-circuit diagnosis method based on combined harmonic spatial current trajectories according to claim 1, characterized in that, The linear transformation process in step S2 satisfies the following relationship: in, Representing the corresponding values ​​in the 3rd harmonic subspace The current component of the phase; These represent the initial phase angles of the currents in phases a, b, c, d, e, and f, respectively. These represent the current components in two orthogonal directions in the third harmonic subspace; Representing the corresponding values ​​in the 5th harmonic subspace The current component of the phase; These represent the current components in two orthogonal directions in the 5th harmonic subspace.

4. The multiphase motor open-circuit diagnosis method based on combined harmonic spatial current trajectories according to claim 1, characterized in that, Step S3 includes the following sub-steps: S31 performs a check and sign reversal operation on the value of each current variable generated in step S2 for the two sets of current variables. That is, if the value of the current variable is negative, it is converted to positive; if the value of the current variable is positive, it remains unchanged, thus completing the inversion of all current variables. S32 determines the window length for the moving average calculation. This window length is selected based on the sampling frequency of the multiphase motor and the dynamic change characteristics of the current signal to ensure that the fluctuation of the current signal can be effectively smoothed without losing fault characteristic information. S33 will sequentially extract continuous sampling data from each group of inverted current variables according to the set window length, and calculate the arithmetic average of the sampling data in each window. S34 moves the window sequentially according to the sampling time order and repeats the arithmetic mean calculation process to obtain two sets of moving average results with the same number of sampling points as the original current variable. Each set of moving average results corresponds to the moving average data of the harmonic current of one phase.

5. The multiphase motor open-circuit diagnosis method based on combined harmonic spatial current trajectories according to claim 1, characterized in that, Step S4 includes the following sub-steps: S41 performs a one-to-one correspondence analysis on the two sets of moving average results obtained in step S3, and uses the two sets of moving average results of the same phase as the horizontal and vertical coordinates of the coordinate points to construct the combined harmonic spatial current trajectory of each phase of the multiphase motor. S42 observes the constructed combined harmonic spatial current trajectory shape. If the trajectory of a certain phase is a fixed single point, that is, the two sets of moving average results of the phase remain constant and unchanged at continuous sampling time. S43 sets a threshold for avoiding the influence of single-point trajectory. This threshold is determined based on the fluctuation range of harmonic current during normal operation of the multiphase motor and the change amplitude of harmonic current during faults. When S44 detects that a certain phase trajectory is a single point, it adjusts the ratio of the two sets of moving average results of that phase according to the set threshold, so that the adjusted ratio is no longer within the preset ratio range corresponding to the 0°, 45°, and 63.43° straight trajectories, thus avoiding fault judgment deviation caused by single-point trajectories.

6. The multiphase motor open-circuit diagnosis method with combined harmonic spatial current trajectory according to claim 1, characterized in that, Step S5 The process includes the following steps: S51 Based on the number of phases of the multiphase motor, define a judgment factor with the same number of phases as the number of phases. Each judgment factor corresponds to one phase winding of the motor, and clarify the relationship between each judgment factor and the corresponding phase winding; S52 Obtain the two sets of moving average results obtained in step S3, and determine the two sets of moving average data corresponding to each judgment factor, namely the 5th harmonic subspace moving average result and the 3rd harmonic subspace moving average result of the phase corresponding to a certain judgment factor; S53 uses division to calculate the value of each judgment factor, taking the 5th harmonic subspace moving average of the corresponding phase as the dividend and the 3rd harmonic subspace moving average of the corresponding phase as the divisor, and calculates the ratio; S54 checks the calculated judgment factor values. If the judgment factor becomes infinite due to the divisor being 0, the judgment factor is assigned a value according to the preset rules to ensure that each judgment factor has a valid value for fault diagnosis.

7. The multiphase motor open-circuit diagnosis method with combined harmonic spatial current trajectory according to claim 1, characterized in that, The multiphase motor is a six-phase open-winding permanent magnet synchronous motor, adopting a six-phase H-bridge topology with a common DC bus. Each phase winding is powered by an H-bridge consisting of four IGBTs. The multiphase winding current collected in step S1 is the six-phase winding current. The two sets of current variables generated in step S2 each have six variables. The number of judgment factors defined in step S5 is six. The open-circuit fault diagnosis table preset in step S6 includes judgment factor features corresponding to 21 types of open-circuit faults, including single-phase open circuit, open circuit between two adjacent phases, open circuit between two phases, and open circuit between two phases.

Images