A method and system for detecting faults of frequency converter

Abstract

The present application relates to the technical field of measuring electrical variables, and specifically to a fault detection method and system for a frequency converter, the method comprising: obtaining the A-phase current data, temperature data and torque data of the frequency converter at each acquisition moment; obtaining each current subsequence, determining the period difference of the current based on the discrete degree of the distribution of the current data in all current subsequences, and determining the waveform change value of the current in combination with the difference in the corresponding torque data between each current subsequence and its adjacent current subsequence after screening; respectively determining the fundamental frequency index of the normal current data and the current data to be detected; determining the deviation value of the current; analyzing the speed of temperature rise, the stability when the temperature reaches the maximum value, and whether the rapid temperature rise is caused by load changes, determining the abnormal temperature rise value; and performing fault detection on the frequency converter. The present application aims to improve the accuracy of open circuit fault detection on the frequency converter.

Description

Technical Field

[0001] The present application relates to the technical field of measuring electrical variables, and in particular to a method and system for detecting faults of a frequency converter. Background Art

[0002] In modern industrial production, frequency converters are used more and more widely as key control equipment for motors in process systems such as ventilation, refrigeration, etc. Frequency converters use insulated-gate bipolar transistors (IGBTs) as switching elements. When the frequency converter works for too long, the IGBTs are prone to heat up. When the heat is severe, the IGBTs will burn out, resulting in open circuit failures. When an IGBT has an open circuit failure, a larger current will flow through the remaining IGBTs, which is very likely to cause overcurrent failures, resulting in more serious accidents.

[0003] Existing methods for detecting open-circuit faults in inverters include the Park vector method, the three-phase average current method, etc. However, these methods all use a single parameter to determine whether an open-circuit fault has occurred, which can easily lead to misjudgment and result in low detection accuracy for open-circuit faults in inverters. Summary of the invention

[0004] In view of the above, it is necessary to provide a method and system for detecting faults of a frequency converter, which improves the accuracy of detecting open-circuit faults of the frequency converter compared with the traditional method for detecting faults of the frequency converter:

[0005] In a first aspect, an embodiment of the present application provides a method for detecting a fault of a frequency converter, the method comprising the following steps:

[0006] Obtain the A-phase current data, temperature data and torque data of the inverter at each acquisition moment;

[0007] Dividing all current data into current subsequences based on peak values ​​in all current data, and determining the period difference of the current based on the discrete degree of distribution of the current data in all current subsequences;

[0008] Screening each current subsequence according to the matching degree between the distribution of each current subsequence and the sinusoidal waveform, and determining the waveform change value of the current based on the difference in corresponding torque data between each current subsequence and its adjacent current subsequence after screening and the period difference;

[0009] Divide all current data into normal current data and current data to be detected, and obtain low-frequency and high-frequency components of the normal current data and the current data to be detected;

[0010] Based on the correlation between the low-frequency and high-frequency components and the normal current data, and in combination with the difference between the low-frequency and high-frequency components, respectively determine the fundamental frequency index of the normal current data and the current data to be detected;

[0011] Determine the deviation value of the current based on the difference between the fundamental frequency index of the normal current data and the current data to be detected and in combination with the waveform change value;

[0012] Analyze the difference between the changing trend of local temperature data and the overall temperature data, the discreteness of the temperature data distribution, and the overlap between the temperature data with an upward trend and the maximum local average level and the torque data with the maximum local average level in the acquisition time among all the temperature data, and determine the abnormal temperature rise value;

[0013] The inverter is subjected to fault detection based on the deviation value and the abnormal temperature rise value.

[0014] In one embodiment, the process of determining the period difference is: calculating the mean value of each current subsequence, and taking the average value of the mean values ​​of all current subsequences and the fusion result of the discreteness as the period difference of the current.

[0015] In one embodiment, the process of determining the waveform change value is as follows:

[0016] According to the acquisition time corresponding to each current subsequence, all torque data are divided into each torque local sequence;

[0017] For each current subsequence whose mean is not 0, the difference between the mean of the torque local sequence corresponding to each current subsequence and its adjacent previous torque local sequence is calculated and recorded as the torque difference;

[0018] Count the number of 0s in each current subsequence, recorded as the number of 0 values, and calculate the average of the number of 0 values ​​of all current subsequences whose mean values ​​are 0 and all current subsequences whose mean values ​​are not 0, recorded as the number of periodic zero values ​​and the number of fluctuating zero values;

[0019] The expression of the waveform change value is:

[0020] ; In the formula, A is the waveform change value of the current; B is the period difference of the current; C is the mean of the torque difference of all current subsequences whose means are not 0; D is the difference between the number of fluctuation zero values ​​and the number of period zero values; τ is a preset value greater than 0.

[0021] In one embodiment, the process of dividing the normal current data and the current data to be detected is as follows:

[0022] Obtaining a segmentation threshold of the mean values ​​of all current subsequences, and obtaining the first data in the current subsequence to which the mean value equals the segmentation threshold value belongs;

[0023] The current data at all sampling moments before the sampling moment of the first data is taken as normal current data;

[0024] The remaining current data is used as current data to be detected.

[0025] In one embodiment, the process of determining the low and high frequency components is as follows:

[0026] Arrange all normal current data in time sequence to form a normal current subsequence, and obtain each inherent mode function of the normal current subsequence; use the last preset number of inherent mode functions as low-frequency components of the normal current data; and use the first preset number of inherent mode functions as high-frequency components of the normal current data;

[0027] For the current data to be detected, the same method as that for obtaining the low and high frequency components of the normal current data is adopted to obtain the low and high frequency components of the current data to be detected.

[0028] In one embodiment, the fundamental frequency index of the normal current data and the current data to be detected is determined as follows:

[0029] The amplitudes of all acquisition moments of each high-frequency component and each low-frequency component of the normal current data are arranged in time sequence to form each high-frequency component sequence and each low-frequency component sequence;

[0030] The mean of the absolute values ​​of the correlations between all low-frequency component sequences and the normal current subsequence is recorded as the low-frequency similarity value;

[0031] The average of the absolute values ​​of the correlations between all high-frequency component sequences and the normal current subsequence is recorded as the high-frequency similarity value;

[0032] The difference between the low-frequency similarity value and the high-frequency similarity value is recorded as the fundamental frequency similarity value;

[0033] Calculate the ratio of the mean of the sum of squares of all elements in all low-frequency component sequences to the mean of the sum of squares of all elements in all high-frequency component sequences;

[0034] The fundamental frequency index of normal current data is positively correlated with the fundamental frequency similarity value and the ratio respectively;

[0035] According to the data to be detected and its low-frequency and high-frequency components, the fundamental frequency index of the current data to be detected is obtained by adopting the same method as that for obtaining the fundamental frequency index of normal current data.

[0036] In one embodiment, the expression of the deviation value is:

[0037] ; In the formula, K is the deviation value of the current; A is the waveform change value of the current; is the fundamental frequency index of normal current data; is the fundamental frequency index of the current data to be detected.

[0038] In one embodiment, the abnormal temperature rise value is determined as follows:

[0039] Obtain a fitting straight line for all temperature data in each temperature subsequence, and record all temperature subsequences whose fitting straight line slope is greater than 0 as temperature rise subsequences;

[0040] The average of the differences in the slopes of the fitted straight lines between each temperature rise subsequence and all other temperature rise subsequences is recorded as the slope mean; the fusion result of the slope mean of each temperature rise subsequence and the slope of the fitted straight line is taken as the temperature rise difference value of each temperature rise subsequence;

[0041] The temperature subsequence with the largest mean is recorded as the temperature detection subsequence;

[0042] The abnormal temperature rise value is proportional to the discreteness of all temperature data in the temperature detection subsequence and the temperature rise difference value of the adjacent previous temperature rise subsequence of the temperature detection subsequence, and is inversely proportional to the overlap degree.

[0043] In one embodiment, the process of performing fault detection on the frequency converter is:

[0044] The fusion result of the deviation value and the abnormal temperature rise value is used as the fault confidence of the frequency converter;

[0045] If the normalized value of the fault confidence is greater than or equal to a preset confidence threshold, it is determined that an open circuit fault has occurred in the frequency converter; otherwise, it is determined that the frequency converter is normal.

[0046] In a second aspect, an embodiment of the present application further provides a fault detection system for a frequency converter, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, the steps of a fault detection method for a frequency converter described in any one of the above items are implemented.

[0047] This application has at least the following beneficial effects:

[0048] The present application evaluates the possibility of abnormal changes in the current waveform through the periodic characteristics of the current and the changes in the load when the current changes; and then combines the fundamental frequency component in the current data to further evaluate the possibility of the current signal deviating from the normal current signal, thereby improving the accuracy of analyzing the possibility of abnormal current signals. Furthermore, through the possibility of abnormal current signals, the possibility of open circuit faults in the inverter is evaluated, thereby improving the accuracy of detecting open circuit faults in the inverter.

[0049] Furthermore, considering the characteristics that load changes may cause temperature rise, the speed of temperature rise, the stability of the temperature when it reaches the maximum value, and whether the rapid temperature rise is caused by load changes are analyzed, and the possibility that the inverter temperature rise is abnormal is evaluated, which improves the accuracy of analyzing whether the inverter temperature rises abnormally, thereby improving the accuracy of detecting open circuit faults in the inverter;

[0050] Furthermore, by comprehensively considering the abnormal conditions of the current signal and the abnormal conditions of the temperature, it is determined whether an open circuit fault occurs in the inverter, and the possibility of an open circuit fault in the inverter is comprehensively analyzed from multiple aspects, thereby improving the accuracy of open circuit fault detection for the inverter. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 A flowchart of a method for detecting a fault in a frequency converter provided by an embodiment of the present application;

[0053] Figure 2 It is a schematic diagram of the process of determining the waveform change value;

[0054] Figure 3 It is a schematic diagram of the process of obtaining normal current data and current data to be detected;

[0055] Figure 4 It is a schematic diagram of the modal decomposition of the normal current subsequence;

[0056] Figure 5 Schematic diagram of the process of determining the fundamental frequency index of normal current data. DETAILED DESCRIPTION

[0057] In the description of the embodiments of the present application, words such as "exemplary", "or", "for example" and the like are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "or", "for example" and the like is intended to present related concepts in a concrete manner.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art in the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. It should be understood that, unless otherwise specified, " / " means or.

[0059] It should also be noted that the terms "first" and "second" in the present application are used to distinguish similar objects rather than to describe a specific order or sequence.

[0060] The following is a detailed description of a method and system for detecting faults in a frequency converter provided by the present application in conjunction with the accompanying drawings.

[0061] See also Figure 1 , which shows a flowchart of a method for detecting a fault of a frequency converter provided by an embodiment of the present application, the method comprising the following steps:

[0062] Step 1: Obtain the A-phase current data, temperature data and torque data of the inverter at each collection time.

[0063] The present application collects the A-phase current data in the output current of the inverter through a current sensor installed on the output end of the inverter; collects the temperature data of the inverter through a temperature sensor installed on the inverter; and obtains the torque data of the fan through a torque sensor installed on the fan drive shaft, so as to reflect whether the load has changed through the torque data. If the torque has been relatively stable and continues to be stable after a sudden change, it can reflect that the fan load has changed to a certain extent.

[0064] All collected A-phase current data, temperature data, and torque data within the preset time length before the current moment are arranged in time sequence to form a current sequence, a temperature sequence, and a torque sequence. In order to eliminate the dimensional influence between the data, the current sequence, the temperature sequence, and the torque sequence are normalized respectively.

[0065] In this embodiment, when collecting A-phase current data, temperature data and torque data, the collection time interval and the preset time length are 1 second and 1 hour respectively. The values ​​of the collection time interval and the preset time length are preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0066] In this embodiment, the Z-Score normalization method is used to normalize the current sequence, temperature sequence, and torque sequence. As other implementation methods, on the basis of being able to realize the normalization of the current sequence, temperature sequence, and torque sequence, the implementer may use other existing technologies for normalization, such as the Min-Max normalization method, the decimal calibration normalization method, etc., and this application does not make any special restrictions.

[0067] Step 2, determine the waveform change value of the current to evaluate whether the current waveform deviates from the waveform shape under the normal state; further combine the fundamental frequency component in the current signal to analyze the possibility of the current signal deviating from the normal current signal; determine the abnormal temperature rise value to evaluate whether the inverter temperature rise is normal temperature rise or abnormal temperature rise caused by load change.

[0068] Under the condition of stable load, if the IGBT of the inverter does not have an open circuit fault, the data in the collected current sequence has a strong periodicity, and the harmonic components in the current signal are small. The temperature of the inverter is normal. Even if the load changes suddenly, the data in the current sequence still has a certain periodicity under the new load state, and only its amplitude will change. If an open circuit fault occurs, the data in the current sequence will no longer have periodicity, and the harmonic components in the current signal will increase. At the same time, when an open circuit fault occurs due to excessive temperature, the temperature of the inverter will also rise.

[0069] Since there are multiple switch tubes in a frequency converter, this embodiment analyzes the open circuit fault of tube T4.

[0070] Step 2.1, dividing all current data into current subsequences based on peak values ​​in all current data, and determining the period difference of the current based on the discrete degree of the distribution of the current data in all current subsequences.

[0071] When the inverter is running without fault, the three-phase current is sinusoidal, and the average value of the current in one cycle is 0; however, after an open circuit fault occurs in a switch tube in the inverter, the average value of the current in the faulty phase in one cycle is no longer 0. After an open circuit fault occurs, the open circuit fault prevents the current from flowing in the negative half cycle, and the current will only remain in the positive half cycle, that is, all current data are greater than or equal to 0.

[0072] A peak-to-valley detection algorithm is used to obtain all peak values ​​in the current sequence, and the current sequence is divided into multiple current subsequences according to the peak values, specifically: the current data between two adjacent peak values ​​form a current subsequence. The discrete degree of the mean of all current subsequences is recorded as the overall discrete degree, the average of the mean of all current subsequences is recorded as the overall mean, and the fusion result of the overall mean and the overall discrete degree is used as the period difference of the current.

[0073] It should be noted that: the greater the period difference, the greater the difference in the mean values ​​between the current subsequences, and if the mean value of each current subsequence is larger, the greater the difference from the mean value of the normal current subsequence, that is, the value of 0, the greater the possibility of non-periodic changes in the current waveform.

[0074] In this embodiment, the degree of dispersion is the standard deviation. As other implementation methods, based on the degree of uneven distribution of the mean value of the measurable current subsequence, the implementer may use other statistics for measurement, such as variance, coefficient of variation, etc., and this application does not impose any special restrictions.

[0075] It should be understood that fusion refers to combining multiple independent variables in a way that enhances the overall effect, such as an additive relationship, a multiplicative relationship, etc. The implementer may limit it according to the actual situation, and this application does not impose any special restrictions.

[0076] In this embodiment, the sum of the overall mean value and the overall discreteness is taken as the period difference of the current.

[0077] In another embodiment, the product of the overall mean value and the overall dispersion degree is taken as the period difference of the current.

[0078] Step 2.2, screening each current subsequence according to the matching degree between the distribution of each current subsequence and the sinusoidal waveform, and determining the waveform change value of the current based on the difference in corresponding torque data between each current subsequence and its adjacent current subsequence after screening and the period difference.

[0079] Considering that when the load changes, the current will suddenly increase or decrease, which may also cause the mean of some current subsequences to be non-zero, therefore, combined with the load change reflected by the torque sequence, it is calculated whether the load has changed when the mean of the current subsequence is not zero.

[0080] According to the acquisition moments corresponding to the current sub-sequences, the torque sequence is divided into torque local sequences corresponding to the current sub-sequences.

[0081] For each current subsequence whose mean is not 0, the difference between the mean of the torque local sequence corresponding to each current subsequence and its adjacent previous torque local sequence is calculated and recorded as the torque difference. It should be noted that the torque difference of the first current subsequence is the difference between the mean of the torque local sequence corresponding to the first current subsequence and its adjacent next torque local sequence. The smaller the torque difference, the less likely the load has changed.

[0082] In this embodiment, the difference in the means between the torque subsequences is the absolute value of the difference. As other implementation methods, based on the measurable difference in the means between the torque subsequences, the implementer may use other calculation methods to measure the difference in the means between the torque subsequences, such as ratio, square of the difference, etc., and this application does not impose any special restrictions.

[0083] The number of 0s in each current subsequence is counted and recorded as the number of 0 values. The average of the number of 0 values ​​of all current subsequences with a mean of 0 is recorded as the number of periodic zero values; the average of the number of 0 values ​​of all current subsequences with a mean value of not 0 is recorded as the number of fluctuating zero values. Since the current cannot flow in the negative half cycle when a fault occurs, the current data is 0 during the time period of the negative half cycle, and there will be many 0 elements in a single current subsequence. If no fault occurs, the number of 0 elements in a single current subsequence will be very small. Therefore, the greater the difference between the number of fluctuating zero values ​​and the number of periodic zero values, the greater the difference between the current subsequence with a mean value of 0 and the current subsequence with a mean value of not 0.

[0084] Based on the difference between the number of fluctuation zero values ​​and the number of periodic zero values, the periodic difference, and the distribution of the torque difference of all current subsequences whose mean values ​​are not 0, the waveform change value of the current is determined, and the expression is:

[0085] ; Wherein, A is the waveform change value of the current; B is the period difference of the current; C is the mean of the torque difference of all current subsequences whose mean is not 0; D is the difference between the number of fluctuation zero values ​​and the number of period zero values; τ is a preset value greater than 0, the purpose of which is to avoid the denominator being 0. The value of τ is preset manually and can be set by the implementer. In this embodiment, the value of τ is 0.01.

[0086] In this embodiment, the difference between the number of fluctuating zero values ​​and the number of periodic zero values ​​is the absolute value of the difference. As other implementation methods, based on the measurable difference between the number of fluctuating zero values ​​and the number of periodic zero values, the implementer may use other calculation methods for measurement, such as ratio, square of difference, etc. This application does not impose any special restrictions.

[0087] It should be noted that: if the period difference B is larger, it means that the mean difference between the current subsequences is larger, and the larger the mean, the more unstable the periodicity of the current signal; if the mean C of the torque difference of all current subsequences whose means are not 0 is smaller, it means that when the current subsequence whose means are not 0 changes, the possibility of load change is smaller, which means that the deviation of the current mean from 0 is more likely to be an abnormality of the current itself, rather than caused by load changes; if the difference D between the number of fluctuating zero values ​​and the number of periodic zero values ​​is larger, it means that the current data in the current subsequence whose means are not 0 are more likely to be concentrated on the same side of the 0 value, and the current waveform is more likely to deviate from the normal current waveform. Therefore, if the waveform change value A is larger, it means that the current waveform is more likely to deviate from the normal periodicity, and the possibility of an open circuit fault in the inverter is greater. The flow chart for determining the waveform change value is as follows. Figure 2 shown.

[0088] Step 2.3, dividing all current data into normal current data and current data to be detected, and obtaining low-frequency components and high-frequency components of the normal current data and the current data to be detected.

[0089] Under normal operating conditions, the current waveform output by the inverter is relatively smooth and is mainly a fundamental sine waveform. If an open circuit fault occurs, the current waveform will be greatly distorted and the high-order harmonic components will increase significantly. Therefore, by comparing the harmonic components in the data before and after the current fluctuation, it is possible to further evaluate whether the inverter has an open circuit fault.

[0090] Get the segmentation threshold of the mean of all current subsequences, get the mean equal to the segmentation threshold, take the first data in the current subsequence to which the mean belongs as the partitioning data, take all current data before the partitioning data in the current sequence as normal current data, and arrange them in time sequence to form a normal current subsequence; take the remaining current data as the current data to be detected, and arrange them in time sequence to form a current subsequence to be detected. It should be noted that: if the segmentation threshold corresponds to multiple current subsequences, the first data in the current subsequence with the smallest order is taken as the partitioning data. The schematic diagram of the process of obtaining normal current data and current data to be detected is shown in the figure. Figure 3 shown.

[0091] In this embodiment, the Otsu threshold segmentation algorithm is used to obtain the segmentation threshold. As other implementation methods, on the basis of being able to obtain the segmentation threshold, the implementer may adopt other existing technologies to obtain the segmentation threshold, such as global threshold segmentation, iterative threshold segmentation, etc., and this application does not impose any special restrictions.

[0092] The normal current subsequence is modally decomposed to obtain the intrinsic mode functions (IMF). The modal decomposition diagram of the normal current subsequence is shown in Figure 4 As shown in the figure, P is the Z-Score normalization result of the normal current subsequence. Since the decomposed IMFs are arranged in order from high frequency to low frequency, the first number of IMFs preset in the front are all regarded as high-frequency components, and the second number of IMFs preset in the back are all regarded as low-frequency components. The amplitudes of all acquisition moments in each high-frequency component are arranged in time sequence to form each high-frequency component sequence; the amplitudes of all acquisition moments in each low-frequency component are arranged in time sequence to form each low-frequency component sequence. For the current data to be detected, the same acquisition method as that of each high-frequency component sequence and each low-frequency component sequence of the normal current data is adopted to obtain each high-frequency component sequence and each low-frequency component sequence of the current data to be detected.

[0093] In this embodiment, a variational modal decomposition algorithm is used to perform modal decomposition on the normal current subsequence. As other implementation methods, on the basis of being able to perform modal decomposition on the normal current subsequence, the implementer may use other existing technologies to perform modal decomposition on the normal current subsequence, such as a set empirical mode decomposition algorithm, a complementary set empirical mode decomposition method, etc., and this application does not impose any special restrictions.

[0094] In this embodiment, the values ​​of the preset first quantity and the preset second quantity are 3 and 2 respectively. The values ​​of the preset first quantity and the preset second quantity are preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0095] Step 2.4, based on the correlation between the low and high frequency components and the normal current data, and in combination with the difference between the low and high frequency components, respectively determine the fundamental frequency index of the normal current data and the current data to be detected.

[0096] The mean of the absolute values ​​of the correlations between all low-frequency component sequences of normal current data and the normal current subsequence is recorded as the low-frequency similarity value; the mean of the absolute values ​​of the correlations between all high-frequency component sequences of normal current data and the normal current subsequence is recorded as the high-frequency similarity value. The difference between the low-frequency similarity value and the high-frequency similarity value is recorded as the fundamental frequency similarity value. The larger the fundamental frequency similarity value, the higher the similarity between the low-frequency part of the current signal and the original normal signal, indicating that the low-frequency characteristics of the current signal dominate, the current waveform is relatively smooth, the harmonic components are less, and it is close to the normal operating state.

[0097] In this embodiment, the correlation is the Pearson correlation coefficient. As other implementation methods, on the basis of being able to measure the correlation between each low-frequency component sequence and the normal current subsequence, and the correlation between each high-frequency component sequence and the normal current subsequence, the implementer may adopt other existing technologies for measurement, such as the Spearman correlation coefficient and the Kendall rank correlation coefficient, and this application does not impose any special restrictions.

[0098] Calculate the mean of the sum of squares of all elements in all low-frequency component sequences of normal current data, and record it as the low-frequency component mean; calculate the mean of the sum of squares of all elements in all high-frequency component sequences of normal current data, and record it as the high-frequency component mean; record the ratio of the low-frequency component mean to the high-frequency component mean as the fundamental frequency ratio. The larger the fundamental frequency ratio, the greater the energy proportion occupied by the low-frequency component in the entire current signal, and the current signal is closer to the normal operating state dominated by the fundamental frequency.

[0099] Combining the fundamental frequency similarity value and the fundamental frequency proportion, the fundamental frequency index of normal current data is obtained, and the expression is:

[0100] ; In the formula, is the fundamental frequency index of normal current data; f is the fundamental frequency similarity value of normal current data; G is the fundamental frequency proportion of normal current data; It is an exponential function with a natural constant as the base. On the one hand, it is to increase the weight, and on the other hand, it is to avoid f being 0, which affects the calculation.

[0101] It should be noted that if the fundamental frequency index The larger the value is, the closer the normal current data is to the fundamental frequency dominated state, that is, the current signal is smoother and has fewer harmonic components. The flowchart for determining the fundamental frequency index of normal current data is as follows: Figure 5 shown.

[0102] According to the current data to be detected and each high-frequency component sequence and each low-frequency component sequence thereof, the fundamental frequency index of the current data to be detected is obtained in the same method as that of obtaining the fundamental frequency index of normal current data.

[0103] Step 2.5, based on the difference between the fundamental frequency indexes of the normal current data and the current data to be detected, combined with the waveform change value, determine the deviation value of the current.

[0104] Furthermore, based on the difference between the fundamental frequency index of the normal current data and the fundamental frequency index of the current data to be detected, combined with the waveform change value of the current, the deviation value of the current is determined, and the expression is:

[0105] ; In the formula, K is the deviation value of the current; A is the waveform change value of the current; is the fundamental frequency index of normal current data; is the fundamental frequency index of the current data to be detected.

[0106] It should be noted that: the larger the value of A, the greater the possibility of abnormal changes in the current waveform; The larger the value is, the greater the difference between the fundamental frequency component of the current signal corresponding to the current data to be detected and the normal current signal is. The smaller the deviation value K is, the smaller the fundamental frequency component in the current signal corresponding to the current data to be detected is. Therefore, if the deviation value K is larger, it means that the abnormal waveform in the current sequence changes more, and the possibility of deviation from the normal current signal is greater, and the possibility of open circuit fault in the inverter is higher.

[0107] Step 2.6, among all the temperature data, analyze the difference in the changing trend between the local temperature data and the overall temperature data, the degree of discreteness of the temperature data distribution, and the degree of overlap in the acquisition time between the temperature data with an upward trend and the maximum local average level and the torque data with the maximum local average level, and determine the abnormal temperature rise value.

[0108] Furthermore, the current deviation value is obtained by analyzing the abnormality of the data in the current sequence. The larger the deviation value, the greater the possibility that the inverter has an open circuit fault. However, considering the randomness of current fluctuations when the load changes, judging whether the inverter has an open circuit fault only by the deviation value may lead to false detection problems. Therefore, combined with the characteristics of the inverter temperature rising rapidly when the open circuit fault occurs, a comprehensive analysis is conducted to further verify the possibility of the fault.

[0109] However, considering that not only open circuit faults can cause rapid temperature increases, but changes in motor load can also cause temperature changes, it is necessary to analyze whether the temperature increase is normal or abnormal.

[0110] The torque sequence is detected for mutation points, and the mutation points are used as segmentation points to segment the torque sequence into multiple torque subsequences. The torque subsequence with the largest mean is recorded as the target torque subsequence. The time corresponding to the target torque subsequence is the maximum load operation time of the motor. By analyzing whether the time when the temperature is highest corresponds to the maximum load operation time of the motor, it is determined whether the temperature rises normally.

[0111] In this embodiment, the Bernaola Galvan segmentation algorithm is used to detect mutation points in the torque sequence. As other implementation methods, on the basis of being able to detect mutation points in the torque sequence, the implementer may use other existing technologies to detect mutation points in the torque sequence, such as the Pettitt mutation point detection algorithm, the Mann-Kendall mutation point detection algorithm, etc., and this application does not impose any special restrictions.

[0112] If no open circuit fault occurs, when the load is the largest, that is, within the time range corresponding to the target torque subsequence, the change trend of the temperature data in the temperature sequence should be a continuous rise first and then a stable fluctuation. If an open circuit fault occurs, regardless of whether the load is the maximum load, the temperature of the inverter will rise rapidly, and the degree of rise and fluctuation will be greater than that of normal temperature changes.

[0113] For the temperature sequence, the temperature sequence is divided into multiple temperature subsequences according to the segmentation method of the torque sequence, and the temperature subsequence with the largest mean is recorded as the temperature detection subsequence.

[0114] Obtain the fitted straight line of all temperature data in each temperature subsequence, and calculate the slope of each fitted straight line. All temperature subsequences with a fitted straight line slope greater than 0 are recorded as temperature rise subsequences. The calculation of the slope is a well-known technique and will not be described in detail in this application.

[0115] In this embodiment, the least squares straight line fitting method is used to obtain the fitting straight line. As other implementation methods, based on the fitting straight line of the temperature data, the implementer can use other existing technologies to obtain the fitting straight line, such as linear regression analysis, weighted least squares method, etc., and this application does not make any special restrictions.

[0116] The average of the differences in the slopes of the fitted straight lines between each temperature rise subsequence and all other temperature rise subsequences is recorded as the slope average; the fusion result of the slope average of each temperature rise subsequence and the slope of the fitted straight line is taken as the temperature rise difference value of each temperature rise subsequence. The larger the temperature rise difference value, the greater the difference and the faster the rise trend of the temperature data in each temperature rise subsequence is compared with other temperature rise subsequences.

[0117] In this embodiment, the product of the slope mean of each temperature rise subsequence and the slope of the fitting straight line is used as the temperature rise difference value of each temperature rise subsequence.

[0118] In another embodiment, the sum of the slope mean value of each temperature rise subsequence and the slope of the fitting straight line is used as the temperature rise difference value of each temperature rise subsequence.

[0119] The order of all temperature data in the temperature detection subsequence and its adjacent previous temperature rise subsequence in the temperature sequence forms a first order set; the order of all torque data in the target torque subsequence in the torque sequence forms a second order set; the similarity between the first order set and the second order set is calculated. If the similarity is greater, it means that there are more repeated elements between the first order set and the second order set, which means that when the temperature is rising and the average temperature is the highest, the corresponding load is also the maximum load.

[0120] In this embodiment, the similarity between the first order set and the second order set is the Jaccard similarity coefficient. As other implementation methods, on the basis of being able to measure the similarity between the first order set and the second order set, the implementer may use other existing technologies for measurement, such as the Sørensen-Dice coefficient, etc., and this application does not impose any special restrictions.

[0121] Based on the similarity between the first sequence set and the second sequence set, the discreteness of all temperature data in the temperature detection subsequence, and the temperature rise difference value of the adjacent temperature rise subsequence, the abnormal temperature rise value of the temperature sequence is determined, and the expression is:

[0122] ; Wherein, L is the abnormal temperature rise value of the temperature sequence; O is the discreteness of all temperature data in the temperature detection subsequence; U is the temperature rise difference value of the adjacent previous temperature rise subsequence of the temperature detection subsequence; Q is the similarity between the first sequence set and the second sequence set; β is a preset value greater than 0, the purpose is to avoid the denominator being 0, the value of β is preset manually, and the implementer can set it by himself. In this embodiment, the value of β is 0.01.

[0123] In this embodiment, the degree of discreteness of the temperature data is the variance. As other implementation methods, based on the measurable degree of uneven distribution of temperature data, the implementer may use other statistics to measure the degree of uneven distribution of temperature data, such as standard deviation, coefficient of variation, etc. This application does not impose any special restrictions.

[0124] It should be noted that: if the degree of dispersion O is larger, it means that in the temperature subsequence with the largest mean, the possibility of drastic changes in temperature data is greater; if the temperature rise difference value U is larger, it means that the temperature data in the adjacent previous temperature rise subsequence of the temperature detection subsequence rises faster, and the temperature rise rate difference between other temperature rise subsequences is greater; if the similarity Q is smaller, it means that the temperature does not rise rapidly and reach the maximum value due to the increase in load. Therefore, if the abnormal temperature rise value L is larger, it means that the rapid rise in temperature is less likely to be caused by normal load changes, and it is more likely to be caused by the open circuit fault of the IGBT inside the inverter.

[0125] Step 3: Perform fault detection on the frequency converter based on the deviation value and the abnormal temperature rise value.

[0126] The fusion result of the current deviation value and the abnormal temperature rise value of the temperature sequence is used as the fault confidence of the inverter. The greater the fault confidence, the greater the possibility that the inverter has an open circuit fault.

[0127] In this embodiment, the product of the current deviation value and the abnormal temperature rise value of the temperature sequence is used as the fault confidence of the inverter.

[0128] In another embodiment, the sum of the current deviation value and the abnormal temperature rise value of the temperature sequence is used as the fault confidence of the frequency converter.

[0129] Furthermore, the fault confidence is normalized and the expression is: ; In the formula, is the normalized value of the fault confidence of the inverter; exp() is an exponential function with a natural constant as the base; W is the fault confidence of the inverter.

[0130] If the normalized value of the fault confidence is greater than or equal to the preset confidence threshold, it is determined that there is a problem with the health status of the inverter and an open circuit fault has occurred in the IGBT inside the inverter; if the normalized value of the fault confidence is less than the preset confidence threshold, it is determined that the health status of the inverter is good and no open circuit fault has occurred in the IGBT inside the inverter.

[0131] In this embodiment, the value of the preset confidence threshold is 0.5. The value of the preset confidence threshold is preset manually and can be set by the implementer. This application does not impose any special restrictions.

[0132] Based on the same inventive concept as the above method, an embodiment of the present application also provides a fault detection system for an inverter, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above-mentioned methods for fault detection of an inverter when executing the computer program.

[0133] In summary, the present application evaluates the possibility of abnormal changes in the current waveform through the periodic characteristics of the current and the changes in the load when the current changes; and then combines the fundamental frequency component in the current data to further evaluate the possibility of the current signal deviating from the normal current signal, thereby improving the accuracy of analyzing the possibility of abnormal current signals. Furthermore, through the possibility of abnormal current signals, the possibility of open circuit faults in the inverter is evaluated, thereby improving the accuracy of detecting open circuit faults in the inverter.

[0134] Furthermore, considering the characteristics that load changes may cause temperature rise, the speed of temperature rise, the stability of the temperature when it reaches the maximum value, and whether the rapid temperature rise is caused by load changes are analyzed, and the possibility that the inverter temperature rise is abnormal is evaluated, which improves the accuracy of analyzing whether the inverter temperature rises abnormally, thereby improving the accuracy of detecting open circuit faults in the inverter;

[0135] Furthermore, by comprehensively considering the abnormal conditions of the current signal and the abnormal conditions of the temperature, it is determined whether an open circuit fault occurs in the inverter, and the possibility of an open circuit fault in the inverter is comprehensively analyzed from multiple aspects, thereby improving the accuracy of open circuit fault detection for the inverter.

[0136] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

[0137] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the basic features of the present application. Therefore, no matter from which point of view, the above embodiments of the present application should be regarded as exemplary and non-restrictive.

Claims

1. A method for detecting a fault in a frequency converter, characterized in that: The method comprises the following steps: Obtain the A-phase current data, temperature data and torque data of the inverter at each acquisition moment; Dividing all current data into current subsequences based on peak values ​​in all current data, and determining the period difference of the current based on the discrete degree of distribution of the current data in all current subsequences; Screening each current subsequence according to the matching degree between the distribution of each current subsequence and the sinusoidal waveform, and determining the waveform change value of the current based on the difference in corresponding torque data between each current subsequence and its adjacent current subsequence after screening and the period difference; Divide all current data into normal current data and current data to be detected, and obtain low-frequency and high-frequency components of the normal current data and the current data to be detected; Based on the correlation between the low-frequency and high-frequency components and the normal current data, and in combination with the difference between the low-frequency and high-frequency components, respectively determine the fundamental frequency index of the normal current data and the current data to be detected; Determine the deviation value of the current based on the difference between the fundamental frequency index of the normal current data and the current data to be detected and in combination with the waveform change value; Analyze the difference between the changing trend of local temperature data and the overall temperature data, the discreteness of the temperature data distribution, and the overlap between the temperature data with an upward trend and the maximum local average level and the torque data with the maximum local average level in the acquisition time among all the temperature data, and determine the abnormal temperature rise value; Performing fault detection on the frequency converter based on the deviation value and the abnormal temperature rise value; The process of determining the waveform change value is as follows: According to the acquisition time corresponding to each current subsequence, all torque data are divided into each torque local sequence; For each current subsequence whose mean is not 0, the difference between the mean of the torque local sequence corresponding to each current subsequence and its adjacent previous torque local sequence is calculated and recorded as the torque difference; Count the number of 0s in each current subsequence, recorded as the number of 0 values, and calculate the average of the number of 0 values ​​of all current subsequences whose mean values ​​are 0 and all current subsequences whose mean values ​​are not 0, recorded as the number of periodic zero values ​​and the number of fluctuating zero values; The expression of the waveform change value is: In the formula, A is the waveform change value of the current; B is the period difference of the current; C is the mean of the torque difference of all current subsequences whose means are not 0; D is the difference between the number of fluctuation zero values ​​and the number of period zero values; τ is a preset value greater than 0.

2. A fault detection method for a frequency converter as claimed in claim 1, characterized in that: The process of determining the period difference is: calculating the mean value of each current subsequence, and taking the average value of the mean values ​​of all current subsequences and the fusion result of the discreteness as the period difference of the current.

3. A fault detection method for a frequency converter as claimed in claim 1, characterized in that: The division process of the normal current data and the current data to be detected is as follows: Obtaining a segmentation threshold of the mean values ​​of all current subsequences, and obtaining the first data in the current subsequence to which the mean value equals the segmentation threshold value belongs; The current data at all sampling moments before the sampling moment of the first data is taken as normal current data; The remaining current data is used as current data to be detected.

4. A method for detecting a fault of a frequency converter as claimed in claim 1, characterized in that: The determination process of the low and high frequency components is as follows: Arrange all normal current data in time sequence to form a normal current subsequence, and obtain each inherent mode function of the normal current subsequence; use the last preset number of inherent mode functions as low-frequency components of the normal current data; and use the first preset number of inherent mode functions as high-frequency components of the normal current data; For the current data to be detected, the same method as that for obtaining the low and high frequency components of the normal current data is adopted to obtain the low and high frequency components of the current data to be detected.

5. A method for detecting a fault of a frequency converter as claimed in claim 1, characterized in that: The process of determining the fundamental frequency index of the normal current data and the current data to be detected is as follows: The amplitudes of all acquisition moments of each high-frequency component and each low-frequency component of the normal current data are arranged in time sequence to form each high-frequency component sequence and each low-frequency component sequence; The mean of the absolute values ​​of the correlations between all low-frequency component sequences and the normal current subsequence is recorded as the low-frequency similarity value; The average of the absolute values ​​of the correlations between all high-frequency component sequences and the normal current subsequence is recorded as the high-frequency similarity value; The difference between the low-frequency similarity value and the high-frequency similarity value is recorded as the fundamental frequency similarity value; Calculate the ratio of the mean of the sum of squares of all elements in all low-frequency component sequences to the mean of the sum of squares of all elements in all high-frequency component sequences; The fundamental frequency index of normal current data is positively correlated with the fundamental frequency similarity value and the ratio respectively; According to the data to be detected and its low-frequency and high-frequency components, the fundamental frequency index of the current data to be detected is obtained by adopting the same method as that for obtaining the fundamental frequency index of normal current data.

6. A method for detecting a fault of a frequency converter as claimed in claim 1, characterized in that: The expression of the deviation value is: Where K is the deviation value of the current; A is the waveform change value of the current; F z is the fundamental frequency index of normal current data; F j is the fundamental frequency index of the current data to be detected.

7. A method for detecting a fault of a frequency converter as claimed in claim 1, characterized in that: The determination process of the abnormal temperature rise value is as follows: Obtain a fitting straight line for all temperature data in each temperature subsequence, and record all temperature subsequences whose fitting straight line slope is greater than 0 as temperature rise subsequences; The average of the differences in the slopes of the fitting lines between each temperature rise subsequence and all other temperature rise subsequences is recorded as the slope average; The fusion result of the slope mean of each temperature rise subsequence and the slope of the fitting straight line is used as the temperature rise difference value of each temperature rise subsequence; The temperature subsequence with the largest mean is recorded as the temperature detection subsequence; The abnormal temperature rise value is proportional to the discreteness of all temperature data in the temperature detection subsequence and the temperature rise difference value of the adjacent previous temperature rise subsequence of the temperature detection subsequence, and is inversely proportional to the overlap degree.

8. A method for detecting a fault of a frequency converter as claimed in claim 1, characterized in that: The process of detecting the fault of the frequency converter is as follows: The fusion result of the deviation value and the abnormal temperature rise value is used as the fault confidence of the frequency converter; If the normalized value of the fault confidence is greater than or equal to a preset confidence threshold, it is determined that an open circuit fault has occurred in the frequency converter; otherwise, it is determined that the frequency converter is normal.

9. A fault detection system for a frequency converter, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the inverter fault detection method as described in any one of claims 1 to 8 are implemented.

Images