Online insulation monitoring method for 380V low-voltage motor

Abstract

The invention relates to the technical field of insulation monitoring, in particular to an on-line insulation monitoring method for a 380V low-voltage motor, which comprises the following steps: acquiring a current response curve, extracting a peak value and an amplitude difference, screening a stable interval, extracting a voltage switching point and a current guide point, calculating displacement, outputting a pairing list, extracting multi-source data, comparing change and generating a transfer track, and calculating a data intersection output waveform code, analyzing a ratio, judging a fluctuation recording path, and outputting a monitoring result. According to the method, a high-confidence screening standard is constructed by extracting stable current peak and amplitude difference features, the synchronous recognition capability is enhanced by combining time displacement analysis of voltage and current guide change points, multi-source thermoelectric indexes are introduced to carry out continuous period comparison to construct a dynamic behavior track, and a complete response fragment is screened by using a multi-dimensional intersection. Fluctuation path tracking and abnormal offset recording are realized through ratio calculation, and the recognition precision and trend judgment capability of weak change of the insulation state are improved.

Description

Technical Field

[0001] This invention relates to the field of insulation monitoring technology, and in particular to an online insulation monitoring method for 380V low-voltage motors. Background Technology

[0002] The field of insulation monitoring technology mainly involves the detection and evaluation of the insulation performance of electrical equipment, especially motors and transformers, under operating conditions. Core aspects include insulation resistance testing, leakage current monitoring, dielectric loss factor analysis, and partial discharge monitoring. Its purpose is to promptly detect potential hazards such as insulation aging, moisture absorption, and breakdown during equipment operation, preventing electrical faults and safety accidents. This technology utilizes online monitoring systems, sensor technology, electrical parameter measurement, and voltage excitation to achieve real-time perception and data acquisition of equipment insulation status. It is widely used in power systems, industrial control, and rail transportation, and is one of the important technologies for ensuring the safe and stable operation of equipment. Traditional online insulation monitoring methods for 380V low-voltage motors refer to the means of testing the insulation performance of 380V AC low-voltage motors under operating conditions. The main problem addressed is that insulation performance is prone to deterioration due to moisture absorption, aging, or temperature rise during long-term motor operation. Traditionally, insulation resistance testing is performed using a megohmmeter during intermittent shutdowns, or low-frequency or pulse voltage is injected online to observe changes in the response current to determine the insulation status.

[0003] Traditional methods suffer from insufficient data continuity during monitoring, are limited by the detection cycle and response mode, lack in-depth analysis of periodic stable fluctuations, struggle to identify subtle early characteristics of insulation state changes, cannot accurately construct temporal correlations between multidimensional physical quantities during motor operation, have low voltage and current signal matching degree leading to weak parameter coupling ability, and lack of multi-index linkage comparison results in blurred state identification boundaries. Especially under conditions of gradual state change or frequent fluctuations, it is difficult to accurately determine insulation trends through static sampling data, which can easily lead to identification bias and false alarms. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a method for online insulation monitoring of 380V low-voltage motors.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an online insulation monitoring method for a 380V low-voltage motor, comprising the following steps:

[0006] S1: Obtain the current response curve, extract the periodic peak value and the amplitude difference between adjacent points, judge the stability based on the power frequency fluctuation reference standard, screen the stable intervals periodically and establish an index, and output the stable band sequence.

[0007] S2: Call the voltage waveform corresponding to the stable band sequence, extract the rising and falling edge switching points, calculate the time displacement with the current-guided change point, filter the delay pairing results, and output the synchronization inflection point pairing list.

[0008] S3: Call the winding temperature rise, dielectric consumption performance, polarization recovery trend and leakage response trend of the synchronous inflection point pairing list during the period covered by the time, compare the continuous changes of the data between adjacent cycles, establish trajectory recursion and determine the direction of change rate offset, and output the insulation thermal transformation transfer trajectory set.

[0009] S4: Call the insulation thermal transformation transfer trajectory set, compare it with the corresponding time period of the stable band sequence and the synchronization inflection point pairing list, calculate the intersection and overlap of the three, and output the continuous response waveform code;

[0010] S5: Call the response curve corresponding to the continuous response waveform code, calculate the offset and hysteresis ratio cycle by cycle, determine whether fluctuation is formed, record the trigger path, and output the online insulation monitoring results of the low-voltage motor.

[0011] As a further embodiment of the present invention, the stable band sequence includes periodic current peak characteristic value, sampling point amplitude difference characteristic, stable interval number index, and power frequency fluctuation reference label; the synchronization inflection point pairing list includes voltage switching point coordinates, current guidance change identifier, time displacement value, and delay pairing result item; the insulation thermal change transfer trajectory set includes winding temperature rise change trajectory, dielectric consumption performance parameters, polarization recovery change characteristics, leakage current response trend trajectory, and change rate offset direction mark; the continuous response waveform code includes complete overlap time interval, data synchronization coverage identifier, and stable sequence mapping number; and the low-voltage motor online insulation monitoring result includes offset expansion ratio, lag time difference ratio, continuous fluctuation path mark, and abnormal trend trigger identifier.

[0012] As a further aspect of the present invention, the stepwise screening of stable intervals refers to screening periodic intervals with stable current response by detecting the amplitude difference between the peak value and adjacent points within each current cycle and judging stability based on the power frequency fluctuation reference standard.

[0013] As a further aspect of the present invention, the screening delay pairing result refers to screening voltage and current inflection point pairing relationships that meet the synchronization requirements based on the time displacement between the voltage rising and falling edges and the current guiding change point.

[0014] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0015] S101: Obtain the current response curve during the online insulation monitoring of the low-voltage motor, extract the peak current of each cycle and the amplitude of adjacent sampling points, calculate the amplitude difference, store the difference data according to the cycle, and obtain the cycle current amplitude difference sequence.

[0016] S102: Based on the cycle current amplitude difference sequence, call the power frequency fluctuation reference standard, determine whether the cycle difference is within the reference range, mark and sort the cycle numbers that meet the conditions, and generate a stable cycle number sequence.

[0017] S103: Based on the stable period number sequence, extract the current waveform segments within the corresponding period, establish a unified index according to the number order and aggregate them into continuous data frames to establish a stable band sequence.

[0018] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0019] S201: Call the voltage waveform data within the corresponding time period of the stable band sequence, and perform edge change trend detection on the voltage signal in each waveform segment, identify the amplitude change positions of the rising and falling edges, record the sampling index and timestamp of the change point in the voltage waveform, and obtain the voltage edge switching point set.

[0020] S202: Based on the voltage edge switching point set and the guiding change point position of the current response curve, calculate the time difference between each pair of switching points and guiding change points to form a set of voltage guiding time difference vectors and generate a current-voltage displacement dataset.

[0021] S203: For the time difference values ​​in the current, voltage and displacement data set, based on the judgment rule that the voltage switching point timestamp is later than the current guiding change point timestamp, filter all time difference data with a delay relationship, extract the corresponding switching point and guiding point pairing index, and establish a synchronous inflection point pairing list.

[0022] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0023] S301: Call the winding temperature rise value, dielectric consumption data, polarization recovery curve and leakage response trend sequence within the coverage period of the synchronization inflection point pairing list, retrieve the continuous data frames of physical quantities in the synchronization period according to the cycle, establish the corresponding time mapping structure, and obtain the insulation response parameter matrix.

[0024] S302: Based on the insulation response parameter matrix, extract the winding temperature rise, dielectric consumption, polarization recovery amplitude and leakage response level of adjacent period segments in the cycle, construct multi-dimensional vector sequences respectively, calculate the change difference of the indicators in adjacent cycles in the order of the cycle, and generate insulation response change sequence group;

[0025] S303: Based on the trend of the index changes in the insulation response change sequence group, determine the direction of change of the index in the continuous period, and perform cumulative offset direction statistics on adjacent differences, aggregate the index change trajectories in the same direction, establish a structured tracking path in time order, and establish an insulation thermal change transfer trajectory set.

[0026] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0027] S401: Call the insulation thermal transformation transfer trajectory set and extract the start and end time index from it. At the same time, retrieve the time period boundary parameters in the pairing list of the stable band sequence and the synchronization inflection point, establish three sets of time interval sets, and obtain the covered time interval mapping set.

[0028] S402: Based on the three sets of time intervals in the covered time interval mapping set, compare the intersection length of the start and end ranges with the interval length segment by segment, calculate the intersection length ratio, and determine whether it is greater than the intersection overlap threshold. Filter the time period numbers that meet the requirements of the intersection ratio and generate an intersection overlap matching list.

[0029] S403: For the number in the intersection coincidence matching list, call the voltage and current waveform sequence within the corresponding time period, generate a unique index code based on the number, bind and identify it with the original signal segment, and establish a continuous response waveform code.

[0030] As a further aspect of the present invention, the specific steps of S5 are as follows:

[0031] S501: Call the continuous response waveform code, retrieve the offset expansion amplitude and lag time difference parameters in each cycle, construct the offset amplitude sequence and lag time sequence according to the cycle, calculate the corresponding ratio, and generate the cycle offset ratio sequence.

[0032] S502: Based on the cycle offset ratio sequence, determine whether the ratio in a continuous cycle remains in a fluctuating increasing or decreasing state. If the continuous fluctuation condition is met, mark the path number for the corresponding cycle position and obtain the continuous fluctuation path index set.

[0033] S503: Based on the path number in the continuous fluctuation path index set, retrieve the associated periodic response curves, ratio sequences and time indexes, aggregate and classify all marked paths, generate identification content according to the number, and establish the online insulation monitoring results of low-voltage motors.

[0034] As a further aspect of the present invention, the intersection overlap threshold refers to the proportion of the intersection of time intervals to the length of any interval, and the proportional standard for verifying the effective synchronization correlation between time periods.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0036] In this invention, a high-confidence screening standard is constructed by extracting the peak value of stable current and the amplitude difference feature. The synchronous identification capability is enhanced by combining the time displacement analysis of voltage and current guidance change points. Multi-source thermoelectric indicators are introduced for continuous periodic comparison to construct dynamic behavior trajectory. Complete response segments are screened by multi-dimensional intersection. Fluctuation path tracking and abnormal offset recording are realized by ratio calculation, thereby improving the identification accuracy and trend judgment capability of weak changes in insulation state. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a schematic diagram of the steps of the present invention;

[0039] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0040] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0041] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0042] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0043] Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0044] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0045] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0046] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0047] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0048] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0049] Please see Figure 1 This invention provides an online insulation monitoring method for a 380V low-voltage motor, comprising the following steps:

[0050] S1: Obtain the current response curve during the online insulation monitoring of the low-voltage motor, extract the amplitude difference between the peak value of the periodic current and the adjacent sampling points, determine whether the difference is stable based on the power frequency fluctuation reference standard, filter the stable interval number periodically and establish an index, and output the stable band sequence.

[0051] S2: Call the voltage waveform of the corresponding time period of the stable band sequence, extract the switching point between the rising edge and the falling edge, calculate the time displacement between the switching point and the current guidance change point, compare the displacement values ​​and filter the pairing results of the delay, and output the synchronization inflection point pairing list.

[0052] S3: Call the synchronous inflection point pairing list to cover the winding temperature rise, dielectric consumption performance, polarization recovery trend and leakage response trend during the period, compare the continuous changes of data between adjacent cycles, establish trajectory recursion and determine the direction of change rate offset, and output the insulation thermal change transfer trajectory set.

[0053] S4: Call the insulation thermal transformation transfer trajectory set, compare it with the corresponding time period of the stable band sequence and the synchronization inflection point pairing list, calculate the intersection and overlap of the three, filter the complete signal segment, and output the continuous response waveform code;

[0054] S5: Call the continuous response waveform code, calculate the ratio of offset expansion to lag time difference cycle by cycle, determine whether a continuous fluctuation trajectory is formed, record the trigger path mark, and output the online insulation monitoring results of the low-voltage motor.

[0055] The stable band sequence includes the peak characteristic value of the periodic current, the amplitude difference characteristic of the sampling point, the stable interval number index, and the power frequency fluctuation reference label. The synchronization inflection point pairing list includes the voltage switching point coordinates, the current guidance change identifier, the time displacement value, and the delay pairing result item. The insulation thermal change transfer trajectory set includes the winding temperature rise change trajectory, dielectric consumption performance parameters, polarization recovery change characteristics, leakage current response trend trajectory, and the change rate offset direction mark. The continuous response waveform code includes the complete overlap time interval, the data synchronization coverage identifier, and the stable sequence mapping number. The online insulation monitoring results of the low-voltage motor include the offset expansion ratio, the lag time difference ratio, the continuous fluctuation path mark, and the abnormal trend trigger identifier.

[0056] Please see Figure 2 The specific steps of S1 are as follows:

[0057] S101: Obtain the current response curve during the online insulation monitoring of the low-voltage motor, extract the peak current of each cycle and the amplitude of adjacent sampling points, calculate the amplitude difference, store the difference data according to the cycle, and obtain the cycle current amplitude difference sequence.

[0058] To obtain the current response curve during online insulation monitoring of a low-voltage motor, under normal energized operation, a current transformer is used to continuously sample the three-phase currents (A, B, and C). With a sampling frequency of 10kHz and a total sampling duration of 2 seconds, 20,000 sampling points can be obtained. Based on the 50Hz power grid frequency, one cycle corresponds to 200 sampling points. First, the entire current data sequence is divided into cycles. Within each cycle, the maximum peak value of the current is extracted. This peak value is obtained by comparing it with the maximum value among the 200 sampling points in the current cycle. Then, for the peak value in each cycle, its preceding value is extracted. The current amplitude at each subsequent sampling point is denoted as follows: previous point amplitude 2.85A, current peak value 3.20A, and subsequent sampling point amplitude 2.75A. The difference between the current peak value and the previous sampling point is calculated to be 0.35A, and the difference between the current peak value and the subsequent sampling point is calculated to be 0.45A. The two differences are then averaged to obtain the current amplitude difference for that cycle as 0.40A. This operation is performed one cycle at a time, and the amplitude differences for all cycles are stored sequentially in a list. For example, the amplitude difference for the first cycle is 0.36A, for the second cycle it is 0.39A, and for the third cycle it is 0.42A, ultimately forming a complete sequence of cycle current amplitude differences.

[0059] S102: Based on the periodic current amplitude difference sequence, call the power frequency fluctuation reference standard, determine whether the period difference is within the reference range, mark and sort the period numbers that meet the conditions, and generate a stable period number sequence.

[0060] To filter the amplitude difference for each cycle, a reference standard needs to be set based on the normal fluctuation range of the current signal at power frequency. This reference standard is set by statistically analyzing the cycle current amplitude difference results from a large number of normal operating samples, taking the average value as the basis for setting a range. For example, if the average amplitude difference measured from 100 samples is 0.42A, and the fluctuation range commonly falls between 0.35A and 0.50A, then the reference standard range is set to 0.35A to 0.50A. When the amplitude difference for a certain cycle falls within this range, it is considered... The current fluctuation in this cycle is in a normal state, and its corresponding cycle number is retained. If the amplitude difference of the 5th cycle is 0.38A, the 6th cycle is 0.42A, and the 7th cycle is 0.47A, then these cycle numbers 5, 6, and 7 all meet the conditions and are recorded as stable cycle numbers. Then, continue to traverse the remaining cycles to finally obtain the sequence of cycle numbers that meet the reference standard. For example, the stable cycle numbers are [5, 6, 7, 12, 13, 14, 20, 21], and arrange them in ascending order to form the final stable cycle number sequence.

[0061] S103: Based on the stable period number sequence, extract the current waveform segments within the corresponding period, establish a unified index according to the number order and aggregate them into continuous data frames to establish a stable band sequence;

[0062] Waveform segments corresponding to each stable period are extracted from the original current response curve. Each period consists of 200 sampling points. For example, the starting sampling point of period 5 is 801 and the ending sampling point is 1000. Therefore, data points 801 to 1000 are extracted as segment 1. For period 6, data points 1001 to 1200 are extracted as segment 2, and so on. All waveform segments are arranged in the original order. After extraction, these segments are reindexed for easier unified management and analysis. The original numbers 5, 6, 7, 12, 13, 14, 20, and 21 are renamed f1 to f8 respectively. Each waveform segment contains three-phase current data with 200 sampling points. After aggregating all segments, a stable band sequence data matrix with a length of 1600 sampling points is obtained, which finally forms a continuous current band sequence with stable period characteristics and indexed by unified numbering.

[0063] Please see Figure 3 The specific steps of S2 are as follows:

[0064] S201: Call the voltage waveform data within the corresponding time period of the stable band sequence, and perform edge change trend detection on the voltage signal in each waveform segment, identify the amplitude change positions of the rising and falling edges, record the sampling index and timestamp of the change point in the voltage waveform, and obtain the voltage edge switching point set.

[0065] The corresponding voltage channel signal is extracted based on the time range consistent with the current sampling. Each cycle has 200 sampling points at a sampling frequency of 10kHz. Voltage data is read point by point, and the sampling time corresponding to each point is recorded. Differential detection is performed on the voltage amplitude in each cycle. The signal change trend is determined by calculating the voltage amplitude change between adjacent sampling points. A threshold of 15V is set for amplitude abrupt change during the detection process. This threshold is derived from long-term collected power frequency waveform samples. Under a 220V RMS voltage system, the voltage peak is approximately 311V, and the voltage difference between normal fluctuation points is generally less than 10V. Therefore, 15V is selected as the abrupt change boundary. When the voltage difference between adjacent sampling points exceeds 15V... The voltage is identified as an edge transition point. For each detection result, the rising or falling trend is determined sequentially. For example, the voltage at point 25 is 112V and at point 26 is 130V, with a difference of 18V and being positive, so it is identified as a rising edge transition point. The voltage at point 46 is 292V and at point 47 is 273V, with a difference of -19V and being negative, so it is identified as a falling edge transition point. The entire waveform data segment is scanned sequentially, and the index of all sampling points that meet the conditions and their corresponding times are recorded in a table. The sampling time interval is 0.1ms, so the time corresponding to point 25 is 2.5ms and the time corresponding to point 47 is 4.7ms. The indices and times of all transition points are arranged and stored in order, and finally a complete set of voltage edge switching points is obtained.

[0066] S202: Based on the voltage edge switching point set and the guiding change point position of the current response curve, calculate the time difference between each pair of switching points and guiding change points, form a set of voltage guiding time difference vectors, and generate a current-voltage displacement dataset.

[0067] The time difference is calculated by comparing and contrasting the data. First, current waveform data within the same time period is read, and the position indices and time records of the guiding change points in the current waveform are compiled into a list. For example, the guiding point indices are 22, 53, 100, 141, 198, 225, etc. The time difference is calculated one by one with each abrupt change point in the voltage edge switching point set. For example, if the voltage abrupt change point time is 12.3ms and the corresponding nearest current guiding point time is 11.9ms, the time difference is 0.4ms; if the voltage point time is 18.7ms and the current point time is 18.5ms, the difference is 0.2ms. The time difference is calculated for all voltage points and guiding points in turn. After each calculation, the result and its respective index number are recorded. Within the 2000 sampling point data segment, about dozens of valid pairing results can be obtained. Each data set consists of voltage point index, current point index, and time difference value. After sorting, a time difference vector data list is formed. Then, all time difference results are sorted in time sequence to generate a current-voltage-displacement dataset, which serves as the data basis for subsequent analysis.

[0068] S203: For the time difference values ​​in the current, voltage and displacement data, based on the judgment rule that the voltage switching point timestamp is later than the current guiding change point timestamp, filter all time difference data with a delay relationship, extract the corresponding switching point and guiding point pairing index, and establish a synchronous inflection point pairing list.

[0069] The process involves item-by-item judgment and filtering based on the following criteria: when the voltage switching point timestamp is later than the current guidance change point timestamp, a delay relationship is considered to exist. During execution, the times of the voltage and current points are read from each group of time difference data and their magnitudes are compared. For example, if a group has a voltage time of 3.1ms and a current time of 2.9ms, the voltage time is later, and this group is retained. If another group has a voltage time of 3.0ms and a current time of 3.2ms, this group is deleted and not included in subsequent statistics. After traversing the entire dataset, all records with delays are retained. During the filtering process, a lower limit of 0.05ms and an upper limit of 1.0ms for delay are set. The ms range is used to remove noise points and insignificant delay data. This range is taken from the statistics of hundreds of normal waveform time difference results, most of which have effective delays between 0.1ms and 0.8ms. Therefore, the above reasonable range is set. After screening, the data retained are, for example, 8 pairs of effective delays. The voltage point index is [25, 56, 102, 145, 201, 230, 278, 300], and the corresponding current point index is [22, 53, 100, 141, 198, 225, 270, 296]. The two correspond one-to-one to form the final synchronization inflection point pairing list, which is output in index form.

[0070] Please see Figure 4 The specific steps of S3 are as follows:

[0071] S301: Call the synchronization inflection point pairing list to cover the winding temperature rise, dielectric consumption data, polarization recovery curve and leakage response trend sequence within the time period, retrieve the continuous data frames of physical quantities in the synchronization time period according to the cycle, establish the corresponding time mapping structure, and obtain the insulation response parameter matrix.

[0072] First, the search range is defined by the start and end time periods of each inflection point in the synchronization inflection point pairing list. The sampling frequency is set to 1kHz, meaning a set of physical quantity data is recorded every 1ms. Each pair of inflection points corresponds to a duration of 4ms. Data from four time points are searched within this range. Physical quantity values ​​are read segment by segment within this time period. For winding temperature rise, a distributed temperature sensor records the time series values. For example, within a certain inflection point time period [120ms~124ms], the temperature rise values ​​are [49.3, 49.6, 49.7, 49.9]. Dielectric consumption values ​​are recorded in real-time by an insulation loss factor detection device. For example, the collected values ​​for this segment are [0.017, 0.018, 0.018, 0.019]. Polarization recovery... The complex voltage was taken from the residual potential value after synchronous application and release voltage, and was collected as [5.0, 5.1, 5.2, 5.3]. The leakage current response value was fed back by the leakage current sensor at the milliampere level, and this data segment was [0.31, 0.33, 0.34, 0.36]. Each type of physical quantity data was labeled according to the sampling time point and formed into a structure. The above operation was repeated for all inflection point pairing time periods. The data segments were summarized to construct a time series dataset of four types of physical quantities with the inflection point pairing index as the primary key. The mapping relationship between each data segment and the start and end time was established. Finally, the response parameter row vector composed of the complete temperature rise, power consumption, polarization voltage and leakage current sequence under each pairing was obtained, and the data were summarized to form the insulation response parameter matrix.

[0073] S302: Based on the insulation response parameter matrix, extract the winding temperature rise, dielectric consumption, polarization recovery amplitude and leakage response level of adjacent period segments in the cycle, construct multi-dimensional vector sequences respectively, calculate the change difference of the indicators in adjacent cycles in the order of the cycle, and generate insulation response change sequence group;

[0074] The average values ​​of various physical quantities within adjacent inflection point period segments are extracted. The average value of the four sampling points for each data segment is calculated as the representative value for that period. For example, if the winding temperature rise data for a certain period segment is [49.3, 49.6, 49.7, 49.9], with an average of 49.625℃; and the next adjacent period is [49.8, 50.1, 50.0, 50.2], with an average of 50.025℃, then the difference in temperature rise between the two periods is +0.4℃. For the dielectric consumption sequences [0.017, 0.018, 0.018, 0.019] and [0.020, 0.021, 0.020, 0.021], the average values ​​are 0.018 and 0.0205 respectively, with a difference of +0.0025. The average values ​​of the polarization recovery curves are [5.0, 5.1, 5.2, 5.3] and [5.4, 5.5, 5.4]. The values ​​of 5.15 and 5.475 are respectively taken as 5.15 and 5.475, with a difference of +0.325; the leakage current responses are [0.31, 0.33, 0.34, 0.36] and [0.35, 0.37, 0.38, 0.39], with average values ​​of 0.335 and 0.3725, with a difference of +0.0375. The difference between the average values ​​of each group of adjacent cycles is calculated to generate the winding temperature rise difference sequence [+0.4, +0.5, -0.2], the dielectric consumption difference sequence [+0.0025, +0.003, -0.0015], the polarization recovery difference [+0.325, +0.2, -0.15], and the leakage current response difference [+0.0375, +0.025, -0.01]. These are arranged in chronological order and combined to form an insulation response change sequence group, which serves as the basis for subsequent trend determination.

[0075] S303: Based on the trend of index changes in the insulation response change sequence group, determine the direction of change of the index in the continuous period, and perform cumulative offset direction statistics on adjacent differences, aggregate the index change trajectories in the same direction, establish a structured tracking path in time order, and establish a set of insulation thermal change transfer trajectories.

[0076] The direction of the difference in each physical quantity within a continuous period is compared item by item. The judgment criteria are set as follows: positive numbers indicate an increase, negative numbers indicate a decrease, and 0 is considered no change. Each physical quantity is marked according to the direction of the difference, and then the cumulative direction of the difference sequence is judged. For example, in the winding temperature rise difference sequence [+0.4, +0.5, -0.2, -0.1, +0.3], the first two items are positive changes, forming a continuous upward trajectory, the middle two items are downward, forming a downward trajectory, and the last item is a new upward segment, which are recorded as trajectory segment 1 (upward), trajectory segment 2 (downward), and trajectory segment 3 (upward). For the dielectric consumption difference sequence [+0.0025, +0.003, -0.0015, +0.002], segment 1 (upward), segment 2 (downward), and segment 3 (downward) are formed. (Decrease), segment 3 (rise); polarization recovery and leakage response are identified and segmented according to the same rules. The start and end times of each segment are derived from the timestamp range of its corresponding inflection point. In this process, a direction switching threshold is set. When the difference between two consecutive cycles is reversed or the absolute value changes by more than 50%, it is judged as a direction switching point. This ratio is based on the actual observation results when the sample trend changes drastically. For example, if the difference in the previous cycle is +0.4 and the difference in the next cycle is -0.2, the change direction is reversed and the amplitude reaches 50%, it is recorded as a switching point. Finally, the continuous cycles in the same direction are merged into a trajectory segment, and the trajectory set sorted by time is output to form an insulation thermal transformation trajectory set. The result includes the start and end cycle number, physical quantity name and offset direction attribute of each trajectory segment.

[0077] Please see Figure 5 The specific steps of S4 are as follows:

[0078] S401: Call the set of insulation thermal transformer transfer trajectories and extract the start and end time indices from them. At the same time, retrieve the time period boundary parameters in the pairing list of stable band sequences and synchronization inflection points, establish three sets of time intervals, and obtain the covered time interval mapping set.

[0079] First, the start and end cycle numbers of each thermal trajectory in the trajectory set are read. The corresponding start and end time indices are calculated using the sampling frequency and cycle length. For example, if a trajectory lasts from cycle 8 to cycle 12, with 200 sampling points per cycle and a sampling frequency of 10kHz, the cycle length is 20ms, the start time is 160ms, and the end time is 240ms. The converted trajectory time interval is [160ms, 240ms]. This process is repeated to extract the time intervals of all trajectories and establish a trajectory time set. Next, each stable cycle number and its time range recorded in the stable band sequence are retrieved. For example, if the stable cycle number is [7, 8, 9, 10, 11], the corresponding time interval is [140ms, 240ms]. [s], organize the stable time set; then read the sampling index of each inflection point in the synchronization inflection point pairing list, for example, the current inflection point is 1320, the voltage inflection point is 1340, the sampling frequency is 10kHz, then the time is 132ms and 134ms respectively. Taking the smaller value as the starting point and the larger value as the ending point, the inflection point interval is obtained as [132ms, 134ms]. Summarize all pairs to form an inflection point time set. After assigning corresponding labels and numbers to the above three types of sets, store them as a unified time interval structure. By comparing and locating the start and end times of the time interval, the construction of the covered time interval mapping set is completed. The result is three types of time interval structure sets containing N segments, M segments and P segments respectively, and the corresponding source and data location can be quickly indexed by number.

[0080] S402: Based on the three sets of time intervals in the coverage time interval mapping set, compare the intersection length of the start and end ranges with the interval length segment by segment, calculate the intersection length ratio, and determine whether it is greater than the intersection overlap threshold. Filter the time period numbers that meet the requirements of the intersection ratio and generate an intersection overlap matching list.

[0081] The specific formula for calculating the intersection length ratio is as follows:

[0082]

[0083] Calculate the intersection ratio R i Each item is checked to see if it meets the set intersection and overlap threshold conditions, and the time period numbers with the required intersection ratio are filtered to generate an intersection and overlap matching list.

[0084] Among them, R i L represents the percentage of intersection of the i-th time intervals. ij L represents the actual intersection length between the i-th time interval and the comparison time interval. it L represents the duration of the i-th time interval itself. ik This represents the effective length of the k-th type of time data within the i-th time interval. This indicates the minimum length of this type of data across all examined time periods. w represents the average length of this type of data across all time periods. k The importance weighting coefficient for the k-th type of time data is represented, with a value ranging from 0 to 1, and m represents the total number of different types of time sets participating in the comparison;

[0085] This formula uses a composite structure, and the molecular part is composed as follows: The first term L ij : Basic intersection length; Second term The weighted sum of the offset increments and weights for comparisons of various time segments; the third item. Suppress the impact of both the intersection and the target duration being too large.

[0086] The denominator consists of two parts: L it : Target segment length; Average deviation of various segments.

[0087] Final result R i Normalized to a dimensionless value.

[0088] Actual parameter values:

[0089] Taking data obtained from actual monitoring tasks as an example, let the target time interval be [120ms, 190ms], the comparison time interval be [150ms, 200ms], and the other parameters be as shown in the table below:

[0090] Table 1 Input Parameters for Intersection Calculation

[0091]

[0092] Substitute into the calculation process

[0093] molecular:

[0094]

[0095] Denominator:

[0096]

[0097] Final result:

[0098]

[0099] The results show that the overlap rate of the i-th time interval with the comparison segment is only 2.18%, which is far below the preset overlap threshold (70%). Therefore, this segment will be removed and will not be included in the sequence of current and voltage response codes in step S403.

[0100] The advantage of the formula lies in the introduction of deviation correction terms for multiple time periods. Nonlinear penalty term Control item for difference from mean It achieves a comprehensive and normalized evaluation of complex time intersection structures, making the calculation results more stable and controllable, and can be widely applied to intersection-dominated time series synchronization problems.

[0101] S403: For the number in the intersection and coincidence matching list, call the voltage and current waveform sequence within the corresponding time period, generate a unique index code based on the number, bind and identify it with the original signal segment, and establish a continuous response waveform code;

[0102] The corresponding original time interval is queried sequentially by number. For example, the time interval corresponding to number IDX_07 is [200ms, 220ms], with a sampling frequency of 10kHz. The converted sampling point range is from point 2000 to point 2200. Then, the waveform sequence data of the voltage channel and current channel are extracted. The data segment corresponding to this time interval is extracted from the full waveform data. The voltage data format is a three-phase channel structure [UA, UB, UC], with 200 points per channel; the current data structure is also a three-phase 200-point structure. After extraction, a unique index is generated for this number. The code name, for example, if the rule is "WRF_number_starting sample value", then the code name is WRF_07_2000. The voltage waveform and current waveform are respectively bound to this code name and recorded in the index mapping table, for example {WRF_07_2000:{voltage data segment, current data segment}}. This mapping structure is used to identify the source, time range and binding position of each waveform with the original sequence, so as to realize subsequent data tracking and matching. Finally, a continuous response waveform code name structure named with WRF code name is established, covering the data segments corresponding to all intersection matching numbers.

[0103] Please see Figure 6 The specific steps of S5 are as follows:

[0104] S501: Call the continuous response waveform code, retrieve the offset expansion amplitude and lag time difference parameters in each period, construct the offset amplitude sequence and lag time sequence according to the period, calculate the corresponding ratio, and generate the period offset ratio sequence.

[0105] For each waveform code, voltage and current response curve data are extracted segment by segment within the corresponding period range. First, the peak voltage and current points for each period are retrieved from the data frame. The offset amplitude between voltage and current is determined by comparing the peak amplitudes. Simultaneously, the sampling times of the peak occurrences are recorded to calculate the lag time difference. Taking a sampling frequency of 10kHz as an example, each sampling point corresponds to a 0.1ms time interval. In the first period, the voltage peak of 310V occurs at point 1480, and the current peak of 297V occurs at point 1496, with a difference of 16 sampling points between the two peaks, a lag time of 1.6ms, and an offset amplitude of 13V. In the second period, the voltage peak is 312V, and the current peak is 295V, a difference of 17V, with a time difference of 1.8ms. The offset in the third period is 15V, with a lag time of 1.4ms; and the offset in the fourth period is 19V, with a lag time of 2.0ms. The offset amplitudes of all periods are recorded sequentially as [13, 17, 15, 19], and the lag time differences are recorded as [1.6, 1.8, 1.4, 2.0]. Positions within the same period are paired one-to-one, and the offset amplitudes and lag times are compared to obtain a ratio sequence, for example [8.1, 9.4, 10.7, 9.5]. A table mapping the ratio sequence to the period numbers is then established. This calculation process is performed across the entire continuous response waveform range, and the output is the complete period offset ratio sequence.

[0106] S502: Based on the cycle offset ratio sequence, determine whether the ratio in a continuous cycle remains in a fluctuating increasing or decreasing state. If the continuous fluctuation condition is met, mark the path number for the corresponding cycle position and obtain the continuous fluctuation path index set.

[0107] The process involves determining whether the fluctuation direction remains consistent across consecutive cycles. First, the difference in ratios between adjacent cycles is calculated; positive values ​​represent an upward trend, and negative values ​​represent a downward trend. Then, it's determined whether the differences across three or more consecutive cycles maintain the same sign. Taking the ratio sequence [8.1, 9.4, 10.7, 9.5, 8.8, 8.2] as an example, the adjacent differences are calculated as [+1.3, +1.3, -1.2, -0.7, -0.6]. It can be seen that the first two segments show a continuous upward trend, while the following three segments show a continuous downward trend. Two conditions must be met simultaneously to determine continuous fluctuation: first, the fluctuation direction must be the same for at least three consecutive cycles; second, the amplitude of a single fluctuation must exceed a set threshold of 0.5. This threshold was derived from analyzing multiple sets of normal motor fluctuation data. Statistical analysis revealed that fluctuations less than 0.3 are mostly random fluctuations, while fluctuations greater than 0.5 reflect significant changes in electrical response. Therefore, 0.5 is used as the judgment threshold. When a continuous cycle segment that meets the conditions is detected, a path marker is added at the corresponding cycle number. For example, cycles 2, 3, and 4 form a continuous upward path, and cycles 4, 5, and 6 form a continuous downward path. These are recorded as path numbers PATH_01 and PATH_02, respectively. All the numbers that meet the conditions are recorded in sequence to form a continuous fluctuation path index set.

[0108] S503: Based on the path number in the continuous fluctuation path index, retrieve the associated periodic response curves, ratio sequences and time indexes, aggregate and classify all marked paths, generate identification content according to the number, and establish online insulation monitoring results for low-voltage motors.

[0109] The system sequentially retrieves the periodic response curves, ratio sequences, and time indices corresponding to each path. First, it reads the start and end period numbers and trend direction of each path, then extracts voltage and current waveform data segments within that range. For example, path number PATH_01 corresponds to periods [2, 3, 4], a time interval of [40ms, 60ms], a ratio sequence of [8.1, 9.4, 10.7], and an increasing trend. Path number PATH_02 corresponds to periods [4, 5, 6], a time interval of [60ms, 80ms], a ratio sequence of [9.5, 8.8, 8.2], and a decreasing trend. Based on the path number, each group of waveform segments is aggregated and organized. Paths with the same trend direction are grouped together; for example, all increasing paths are aggregated into "Path Group A," and decreasing paths are aggregated into "Path Group B." A unique identifier is generated for each path, named according to the rule "MON_Number_Trend," such as MON_01_UP, MON_02_DOWN. The corresponding waveform data, ratio sequence, and time index are then bound and stored in the results file. The final result is a list of monitoring results containing path number, trend attribute, period range, ratio change and time information, which is the online insulation monitoring result of low voltage motor.

[0110] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for online insulation monitoring of a 380V low-voltage motor, characterized in that, Includes the following steps: S1: Obtain the current response curve of a 380V low-voltage motor, extract the periodic peak value and the amplitude difference between adjacent points, judge the stability based on the power frequency fluctuation reference standard, filter and establish an index for each period, and output a stable band sequence. S2: Call the voltage waveform corresponding to the stable band sequence, extract the rising and falling edge switching points, calculate the time displacement with the current-guided change point, filter the delay pairing results, and output the synchronization inflection point pairing list. S3: Call the synchronous inflection point pairing list covering the time period, extract the winding temperature rise, dielectric consumption performance, polarization recovery trend and leakage response trend, compare the continuous changes of data between adjacent cycles, establish trajectory recursion and determine the direction of change rate offset, and output the insulation thermal transformation transfer trajectory set. S4: Call the insulation thermal transformation transfer trajectory set, compare it with the corresponding time period of the stable band sequence and the synchronization inflection point pairing list, calculate the intersection and overlap of the three, and output the continuous response waveform code; S5: Call the response curve corresponding to the continuous response waveform code, calculate the offset and hysteresis ratio cycle by cycle, determine whether fluctuation is formed, record the trigger path, and output the online insulation monitoring results of the low-voltage motor.

2. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The stable band sequence includes periodic current peak characteristic value, sampling point amplitude difference characteristic, stable interval number index, and power frequency fluctuation reference label. The synchronization inflection point pairing list includes voltage switching point coordinates, current guidance change identifier, time displacement value, and delay pairing result item. The insulation thermal change transfer trajectory set includes winding temperature rise change trajectory, dielectric consumption performance parameters, polarization recovery change characteristics, leakage current response trend trajectory, and change rate offset direction mark. The continuous response waveform code includes complete overlap time interval, data synchronization coverage identifier, and stable sequence mapping number. The low-voltage motor online insulation monitoring result includes offset expansion ratio, lag time difference ratio, continuous fluctuation path mark, and abnormal trend trigger identifier.

3. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The term "cycle-by-cycle stable interval screening" refers to screening a periodic interval with stable current response by detecting the amplitude difference between the peak value and adjacent points within each current cycle and judging stability based on the power frequency fluctuation reference standard.

4. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The screening delay pairing result refers to the screening of voltage and current inflection point pairing relationships that meet the synchronization requirements based on the time displacement between the voltage rising and falling edges and the current guiding change point.

5. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Obtain the current response curve during the online insulation monitoring of the low-voltage motor, extract the peak current of each cycle and the amplitude of adjacent sampling points, calculate the amplitude difference, store the difference data according to the cycle, and obtain the cycle current amplitude difference sequence. S102: Based on the cycle current amplitude difference sequence, call the power frequency fluctuation reference standard, determine whether the cycle difference is within the reference range, mark and sort the cycle numbers that meet the conditions, and generate a stable cycle number sequence. S103: Based on the stable period number sequence, extract the current waveform segments within the corresponding period, establish a unified index according to the number order and aggregate them into continuous data frames to establish a stable band sequence.

6. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Call the voltage waveform data within the corresponding time period of the stable band sequence, and perform edge change trend detection on the voltage signal in each waveform segment, identify the amplitude change positions of the rising and falling edges, record the sampling index and timestamp of the change point in the voltage waveform, and obtain the voltage edge switching point set. S202: Based on the voltage edge switching point set and the guiding change point position of the current response curve, calculate the time difference between each pair of switching points and guiding change points to form a set of voltage guiding time difference vectors and generate a current-voltage displacement dataset. S203: For the time difference values ​​in the current, voltage and displacement data set, based on the judgment rule that the voltage switching point timestamp is later than the current guiding change point timestamp, filter all time difference data with a delay relationship, extract the corresponding switching point and guiding point pairing index, and establish a synchronous inflection point pairing list.

7. The online insulation monitoring method for 380V low-voltage motors according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Call the synchronization inflection point pairing list covering the time period, extract the winding temperature rise value, dielectric consumption data, polarization recovery curve and leakage current response trend sequence, retrieve the continuous data frames of physical quantities in the synchronization time period according to the cycle, establish the corresponding time mapping structure, and obtain the insulation response parameter matrix. S302: Based on the insulation response parameter matrix, extract the winding temperature rise, dielectric consumption, polarization recovery amplitude and leakage response level of adjacent period segments in the cycle, construct multi-dimensional vector sequences respectively, calculate the change difference of the indicators in adjacent cycles in the order of the cycle, and generate insulation response change sequence group; S303: Based on the trend of the index changes in the insulation response change sequence group, determine the direction of change of the index in the continuous period, and perform cumulative offset direction statistics on adjacent differences, aggregate the index change trajectories in the same direction, establish a structured tracking path in time order, and establish an insulation thermal change transfer trajectory set.

8. The online insulation monitoring method for a 380V low-voltage motor according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Call the insulation thermal transformation transfer trajectory set and extract the start and end time index from it. At the same time, retrieve the time period boundary parameters in the pairing list of the stable band sequence and the synchronization inflection point, establish three sets of time interval sets, and obtain the covered time interval mapping set. S402: Based on the three sets of time intervals in the covered time interval mapping set, compare the intersection length of the start and end ranges with the interval length segment by segment, calculate the intersection length ratio, and determine whether it is greater than the intersection overlap threshold. Filter the time period numbers that meet the requirements of the intersection ratio and generate an intersection overlap matching list. S403: For the number in the intersection coincidence matching list, call the voltage and current waveform sequence within the corresponding time period, generate a unique index code based on the number, bind and identify it with the original signal segment, and establish a continuous response waveform code.

9. The online insulation monitoring method for a 380V low-voltage motor according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Call the continuous response waveform code, retrieve the offset expansion amplitude and lag time difference parameters in each cycle, construct the offset amplitude sequence and lag time sequence according to the cycle, calculate the corresponding ratio, and generate the cycle offset ratio sequence. S502: Based on the cycle offset ratio sequence, determine whether the ratio in a continuous cycle remains in a fluctuating increasing or decreasing state. If the continuous fluctuation condition is met, mark the path number for the corresponding cycle position and obtain the continuous fluctuation path index set. S503: Based on the path number in the continuous fluctuation path index set, retrieve the associated periodic response curves, ratio sequences and time indexes, aggregate and classify all marked paths, generate identification content according to the number, and establish the online insulation monitoring results of low-voltage motors.

10. The online insulation monitoring method for a 380V low-voltage motor according to claim 8, characterized in that, The intersection overlap threshold refers to the proportion of the intersection of time intervals to the length of any interval, and the proportional standard used to verify the effective synchronization correlation between time periods.

Images