A non-invasive three-phase motor operating state recognition method

Abstract

The application provides a non-invasive three-phase motor operation state recognition method. It includes signal sampling step, pretreatment step, high-frequency feature calculation step, low-frequency event detection step, operation state inspection step and real-time recording step. The data collected by high-precision current and voltage sensor are processed based on annular buffer data structure to obtain differential current cycle. The specific harmonic of current cycle is extracted based on fast Fourier transform algorithm, and high-frequency feature set is constructed by combining other statistical features. The events are detected by calculating the active power and reactive power of the cycle in the annular buffer and judging the extreme points. The events are processed for three-phase motor data source respectively, the state of three-phase motor is judged and recorded, so as to realize comprehensive monitoring of three-phase motor operation state. The application has the advantages that the problem of insufficient comprehensive and accurate monitoring of three-phase motor by the existing three-phase motor operation state recognition method is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power utilization, in particular to a non-intrusive three-phase motor operating state recognition method. BACKGROUND

[0002] Three-phase motors play a crucial role in modern industrial systems and are the core power source for efficient production in many fields. In the manufacturing industry, from stamping, welding, and painting in automobile manufacturing to lathe, milling machine, and grinding machine operations, three-phase motors provide stable and powerful power support for various precision machinery, ensuring the continuity and accuracy of production processes. In the energy exploitation industry, whether it is large-scale oil pumping machines on oil drilling platforms or ventilation and drainage equipment in coal mines, the reliable operation of three-phase motors is crucial to the safety and efficiency of the entire exploitation operation. In construction sites, tower cranes, concrete mixers, and construction elevators rely on three-phase motors for driving.

[0003] Traditional monitoring methods for three-phase motors mainly use invasive detection schemes. This approach requires professional maintenance personnel to open the motor housing under shutdown conditions and use various professional detection instruments such as oscilloscopes, multimeters, etc. to access key components such as stator windings, rotor windings, brushes, commutators (for DC motors or some wound asynchronous motors), and terminal connections. The operation process is not only complex and time-consuming, but also has many drawbacks. For example, frequent disassembly of the motor housing and terminal connections can easily cause mechanical damage to the connection parts, causing bolts to loosen and threads to wear, resulting in increased contact resistance during subsequent operation, which can cause safety hazards such as overheating and sparking. During the process of connecting detection instruments, if not handled carefully, it is easy to scratch the insulation layer inside the motor, exposing the windings, core, and other critical components, significantly increasing the risk of electrical leakage and endangering the safety of the operator. Moreover, since invasive detection is usually done on a regular basis, it is not possible to obtain real-time and continuous motor operating state data, making it difficult to capture some instantaneous or short-lived fault signals, leaving the motor in a "monitoring blind area" between two inspection intervals, which poses a risk to long-term stable operation.

[0004] Nowadays, with the development of industrial automation, intelligent manufacturing and Internet of Things technology, enterprises urgently need to realize non-invasive, high-precision and real-time monitoring of three-phase motors. The present application emerges as the times require, builds a new monitoring system, including data acquisition, preprocessing, high-frequency and low-frequency feature analysis, operation state inspection and event reporting modules. The modules work together to not only ensure the safety and stability of the motor itself, but also provide data support for the intelligent operation and maintenance of industrial production systems, helping to upgrade and sustainably develop the industry. Among them, the data acquisition module uses high-precision sensors to sample current and voltage signals at a safe external position of the motor; the data preprocessing module processes data with the help of a ring buffer; the high-frequency feature calculation module extracts features such as harmonics; the low-frequency event detection module calculates power and judges events; the motor operation state inspection module comprehensively judges the motor opening, closing, open-phase and other states; the motor event reporting module timely cooperates with other devices to report motor events, prompting the system to adjust in time. SUMMARY

[0005] The present application provides a non-invasive three-phase motor operation state recognition method to solve the problem that the existing three-phase motor operation state recognition method is not comprehensive and accurate enough in monitoring three-phase motors.

[0006] The present application provides a non-invasive three-phase motor operation state recognition method, which specifically includes signal sampling, preprocessing, high-frequency feature calculation, low-frequency event detection and operation state inspection steps.

[0007] The signal sampling step is to synchronously sample the current variation and voltage variation of different phase lines of the three-phase motor by a high-precision current-voltage sensor, and convert the collected voltage variation and current variation into voltage signals and current signals; the preprocessing step is to preprocess the current signals and voltage signals based on a ring buffer data structure to obtain a first cycle after difference; the high-frequency feature calculation step is to extract the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude of the three-phase motor, calculate the characteristic zero rate, the U-I trajectory envelope area and the margin of the first cycle, and construct the characteristic zero rate, the U-I trajectory envelope area, the margin, the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude into a high-frequency feature set, mark the first cycle of the three-phase motor which meets the high-frequency feature of the high-frequency feature set to obtain a second cycle; the low-frequency event detection step is to create a ring buffer, calculate the active power and the reactive power in each cycle in the ring buffer, if there is a second cycle in the cycle, mark the data points of the second cycle with high-frequency features, and each position in the ring buffer is used to store the active power, the reactive power and the high-frequency feature mark, if the active power and the reactive power stored in a position meet the preset condition, it is determined that there is a low-frequency event; the running state inspection step is to detect the low-frequency event of each split phase of the three-phase motor, if there is a high-frequency feature, detect the low-frequency event of each split phase of the three-phase motor with low-frequency feature, judge the running state of each split phase according to the result of low-frequency feature detection, and determine the running state of the three-phase motor through the running state of each split phase and the time error of the running state.

[0008] Further, the preprocessing step specifically includes a buffer establishment step, a maximum value alignment step, a Manhattan distance calculation step, a cycle correction step and a difference operation step.

[0009] The buffer establishment step is to set a voltage ring buffer and a current ring buffer, and each ring buffer position stores the data of one cycle; the maximum value alignment step is to find the maximum value index of the head pointer and the tail pointer by traversal, and to align the tail pointer voltage and the head pointer voltage, and the calculation formula is

[0010]

[0011] Among them, V(k) represents the cycle voltage at the head pointer of the aligned voltage ring buffer, k=0, 1,..., M-1, k represents the data point of the cycle, and V h V(k) represents the voltage at the head pointer position of the voltage ring buffer, V(k) represents the voltage at the head pointer position of the voltage ring buffer, M represents the maximum index of the voltage ring buffer head pointer, M = 32.

[0012] The Manhattan distance calculation step sets an offset, traverses all offsets to calculate the Manhattan distance between the frequency at the voltage ring buffer head pointer and the frequency at the voltage ring buffer tail pointer, and records the minimum distance d min and the corresponding offset s min The calculation formula is

[0013]

[0014] Where d(s) represents the Manhattan distance, s represents the offset, s = -5, -4,..., 4, 5, V t represents the voltage at the voltage ring buffer tail pointer position.

[0015] The frequency correction step corrects the frequency of the current ring buffer head pointer according to the offset s min calculated in the voltage ring buffer, and the formula is

[0016]

[0017] Where, represents the corrected frequency current at the current ring buffer head pointer, I h represents the current at the current ring buffer head pointer.

[0018] The difference operation step is to perform a difference operation on the corrected current frequency at the current ring buffer head pointer and the electric frequency at the current ring buffer tail pointer to obtain the differential current frequency, which is the first frequency, and the formula is

[0019]

[0020] Where, I diff represents the differential current frequency, I t represents the current at the current ring buffer tail pointer.

[0021] Further, the high-frequency feature calculation step specifically includes a harmonic amplitude extraction step, a zero rate self-definition step, an envelope area calculation step, and a margin calculation step.

[0022] The harmonic amplitude extraction step extracts the first, second, and third harmonic amplitudes of the three-phase motor through a discrete Fourier transform formula, and the discrete Fourier transform formula is

[0023]

[0024] wherein f=0,1,2.....N-1, f represents the frequency generated when the motor is running, N represents the length of the cycle, W represents the complex matrix, n represents the index of the cycle length, and represents the position in the cycle.

[0025] The zero rate self-defining step is to define the proportion of the values near the zero point in the cycle current by the zero rate, and the formula is

[0026]

[0027] wherein count low represents the number of values near the zero point, n represents the number of cycles, n=32, I i represents the cycle current, lowbound represents the upper threshold of the definition of the value near the zero point, and minorrate represents the zero rate.

[0028] The envelope area calculation step is to calculate the envelope area of the relationship curve between the voltage and the current of each phase of the three-phase motor, and the formula is

[0029]

[0030] wherein S U-I represents the envelope area, n represents the number of cycles, n=32, I i represents the cycle current, U i represents the cycle voltage.

[0031] The margin calculation step is to calculate the sharpness of the current peak in the waveform, and the formula is

[0032]

[0033] wherein x r represents the square root amplitude of the current, n represents the number of cycles, n=32, I i represents the cycle current, x peak represents the current amplitude, and C e represents the calculated margin.

[0034] Further, the judgment standard whether the high-frequency feature set meets the high-frequency feature of the three-phase motor is that the first harmonic amplitude is greater than a first amplitude threshold, the second harmonic amplitude is greater than a second amplitude threshold, the third harmonic amplitude is greater than a third amplitude threshold, the zero rate is less than a zero rate threshold, the U-I trajectory envelope area is within a preset area threshold range, and the margin is within a preset margin threshold range.

[0035] Further, the low frequency event detecting step specifically comprises an active power calculating step, a reactive power calculating step, an active power cumulative sum calculating step and a reactive power cumulative sum calculating step.

[0036] The active power calculating step is used to calculate the active power of each cycle in the ring buffer, and the formula is

[0037] P = UI cosφ

[0038] Wherein, P represents the active power, U represents the cycle voltage, and I represents the cycle current.

[0039] The reactive power calculating step is used to calculate the reactive power of each cycle in the ring buffer, and the formula is

[0040] Q = UI sinφ

[0041] Wherein, Q represents the reactive power.

[0042] The active power cumulative sum calculating step is used to calculate the cumulative sum of the active power of each cycle in the ring buffer, and the formula is

[0043]

[0044] Wherein, i represents the time, P k represents the active power at time k, P0 represents the initial value of the active power, S p (i) represents the cumulative sum of the active power.

[0045] The reactive power cumulative sum calculating step is used to calculate the cumulative sum of the reactive power of each cycle in the ring buffer, and the formula is

[0046]

[0047] Wherein, Q k represents the reactive power at time k, Q0 represents the initial value of the reactive power, S Q (i) represents the cumulative sum of the reactive power.

[0048] Further, the preset condition is that if there is a position m that satisfies the following two preset conditions, it is determined that the mth position and the 0th position constitute a low frequency event:

[0049]

[0050] Wherein, 1≤m≤n, abs(S P (m)) represents the cumulative sum of the data points of the active power, and abs(S Q (m)) represents the cumulative sum of the data points of the reactive power.

[0051] Furthermore, the operation status inspection step specifically includes a high-frequency feature inspection step, a low-frequency feature inspection step and a comprehensive judgment step.

[0052] The high-frequency feature inspection step is to detect whether the low-frequency event has high-frequency features for each phase-detected low-frequency event. The data point set included in the low-frequency event is E={e1, e2, ..., e n}, where n represents the number of data points contained in the event, e i represents the i-th data point, i=1,2,...,n; define the labeling function M(e i ), used to determine the data point e i Is it marked? When the data point e i When marked, M(e i )=1, and execute the next step; when the data point e i When not marked, M(e i )=0, and detect the next low-frequency event; the low-frequency feature inspection step is to define the low-frequency event that passes the high-frequency feature inspection step as a marked low-frequency event, and determine whether the marked low-frequency event has a transient low-frequency feature of motor startup. If so, and the reactive power of the marked low-frequency event is within the power threshold range, it is determined that a motor event is detected in each phase; the comprehensive determination step is based on the motor event detected in each phase, and determines whether the three-phase motor is in one of the three states of three-phase motor on, three-phase motor missing phase and three-phase motor off.

[0053] Furthermore, the low-frequency feature inspection step specifically includes a peak detection step, a peak ratio detection step and a power jump detection step.

[0054] The peak detection step is to detect the first low-frequency event E1 = {e 11 ,e 12 ,...,e 1n}The next second adjacent low-frequency event E2 = {e 21 ,e 22 ,...,e 2n} Whether the following detection conditions are met, if so, execute the next step, if not, detect the next marked low-frequency event pair, the detection conditions are

[0055]

[0056] e 21 -e 1n ==1

[0057] Among them, P represents active power and Q represents reactive power.

[0058] The peak variation ratio detection step is used to determine whether the peak variation ratio meets the determination condition, and if so, the next step is performed, and if not, the next peak variation ratio is detected, and the determination condition is

[0059]

[0060] 1.2 < g < 2.1

[0061] Wherein, g represents the peak variation ratio.

[0062] The power jump detection step is to determine whether the reactive power of the low-frequency event is within the power threshold range by merging the first marked low-frequency event E1 = {e 11 ,e 12 ,...,e 1n} and the second marked low-frequency event E2 = {e 21 ,e 22 ,...,e 2n} reactive power change, the reactive power of the low-frequency event is calculated by the formula

[0063] qdiff = Q[e 2n ]-Q[e 11 ]

[0064] Wherein, qdiff represents the reactive power of the low-frequency event, when 50 < qdiff < 300, it is determined that the reactive power of the low-frequency event is within the power threshold range, and there is a motor start event.

[0065] Further, the comprehensive determination step specifically includes a three-phase motor start determination step, a three-phase motor open-phase determination step, and a three-phase motor shutdown determination step.

[0066] The three-phase motor start determination step is to calculate the maximum time deviation of the three phase motor start event occurrence time when each phase detects a motor start event, and the formula is

[0067] Δt max (t1,t2,t3) = max(|t1-t2|,|t2-t3|,|t1-t3|)

[0068] Wherein, Δt max represents the maximum time deviation, t1, t2, t3 respectively represent the motor start event occurrence time, when Δt max <1, it is determined that the three-phase motor is in the start state.

[0069] The three-phase motor open-phase determination step is that, on the premise of detecting a three-phase motor open event, when a split phase detects a motor closed event within a preset time period, and other split phases have motors running, it is determined that the three-phase motor is in an open-phase state.

[0070] The three-phase motor closed determination step is that, when no motor open event is detected by each split phase, it is determined that the three-phase motor is in a closed state.

[0071] Further, the running state verification step further includes a real-time recording step, for the detected three-phase motor, it is matched by a unique number, and the running state of the three-phase motor corresponding to the number is recorded in real time.

[0072] The application provides a non-invasive three-phase motor running state recognition method, which adopts a high-precision current and voltage sensor to synchronously sample the current and voltage signals of different phase lines of the three-phase motor at a sampling frequency of not less than 1600 Hz, processes the data collected by the high-precision current and voltage sensor based on a ring buffer data structure to obtain the differential current cycle, extracts specific harmonics of the current cycle based on a fast Fourier transform algorithm, constructs a high-frequency feature set in combination with other statistical characteristics, calculates the active power and reactive power of the cycle in the ring buffer, and judges the extreme points to detect events, processes the events on the three-phase motor data source, judges the state of the three-phase motor, and records it, so as to realize comprehensive monitoring of the running state of the three-phase motor, and solve the problem that the existing three-phase motor running state recognition method is not comprehensive and accurate enough in monitoring the three-phase motor. BRIEF DESCRIPTION OF DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0074] Figure 1 is a flowchart of the non-invasive three-phase motor running state recognition method described in the embodiments of the application;

[0075] Figure 2 is a flowchart of the preprocessing step described in the embodiments of the application;

[0076] Figure 3 is a flowchart of the high-frequency feature calculation step described in the embodiments of the application;

[0077] Figure 4 is a flowchart of the low-frequency event detection step described in the embodiments of the application;

[0078] Figure 5 is a flow chart of the running state verification step described in the embodiments of the present application;

[0079] Figure 6 is a flow chart of the low-frequency feature verification step described in the embodiments of the present application;

[0080] Figure 7 is a flow chart of the comprehensive determination step described in the embodiments of the present application. DETAILED DESCRIPTION

[0081] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0082] As shown in Figure 1 , the present application provides a non-invasive three-phase motor running state recognition method, which specifically comprises the following steps: S1) signal sampling step, S2) preprocessing step, S3) high-frequency feature calculation step, S4) low-frequency event detection step, S5) running state verification step, and S6) real-time recording step.

[0083] Step S1) signal sampling step: the current and voltage changes of different phase lines of the three-phase motor are synchronously sampled by a high-precision current and voltage sensor, and the collected voltage and current changes are converted into voltage and current signals.

[0084] In the present embodiment, a high-precision current and voltage sensor is used to synchronously sample the current and voltage signals of different phase lines of the three-phase motor at a sampling frequency of 1600 Hz. The selection of this sampling frequency is based on in-depth research on the running characteristics and fault signal frequency range of the three-phase motor. The current and voltage signals of the three-phase motor will change in different frequency components under normal operation and fault conditions. The 1600 Hz can ensure that sufficient fine signal details are captured, covering both the fundamental frequency information of the three-phase motor in steady state operation and the high-frequency harmonic components caused by faults. The sensor is installed in a safe position outside the three-phase motor close to the wiring terminal, and cleverly uses the principle of electromagnetic induction to non-invasively perceive the current and voltage changes, avoiding any interference to the internal structure of the three-phase motor. Through the sensor, these time-varying physical quantities are converted into electrical signals and transmitted to the subsequent processing unit for subsequent steps.

[0085] Step S2) preprocessing step: the current and voltage signals are preprocessed based on a ring buffer data structure to obtain the first cycle after difference.

[0086] In the embodiment, considering the continuity and real-time processing requirements of motor operation data, based on the ring buffer data structure, the current ring buffer and the voltage ring buffer are defined. First, the maximum value of the cycle at the head and tail pointers of the voltage ring buffer is aligned, and then a cycle is fixed. The Manhattan distance is calculated by continuously offsetting another cycle from the fixed cycle, and the minimum distance and the offset are recorded. The cycle of the head and tail pointers in the current ring buffer is corrected according to the voltage offset, and the difference between the cycle at the tail pointer and the cycle at the head pointer is processed to obtain the differential cycle.

[0087] As shown in Figure 2 , the preprocessing step S2) specifically includes a buffer establishment step S21), a maximum value alignment step S22), a Manhattan distance calculation step S23), a cycle correction step S24), and a difference operation step S25).

[0088] The buffer establishment step S21) sets the voltage ring buffer and the current ring buffer. The length of the voltage ring buffer and the current ring buffer is N=10, and each position of the ring buffer stores one cycle of data. Each cycle of data includes 32 data points.

[0089] The maximum value alignment step S22) finds the maximum value index of the head pointer and the tail pointer by traversal, and aligns the voltage at the tail pointer with the voltage at the head pointer. The calculation formula is

[0090]

[0091] wherein, represents the voltage at the head pointer of the aligned voltage ring buffer, k=0, 1,..., M-1, k represents the data point of the cycle, and V h represents the voltage at the head pointer position of the voltage ring buffer, represents the maximum value index of the tail pointer of the voltage ring buffer, represents the maximum value index of the head pointer of the voltage ring buffer, and M=32.

[0092] The Manhattan distance calculation step S23) fixes the cycle at the tail pointer of the voltage ring buffer, sets the offset, traverses all offsets to calculate the Manhattan distance between the cycle at the head pointer of the voltage ring buffer and the cycle at the tail pointer of the voltage ring buffer, and records the minimum distance d min and the corresponding offset s min , and the calculation formula is

[0093]

[0094] wherein d(s) represents the Manhattan distance, s represents the offset, s = -5, -4,..., 4, 5, V t represents the voltage at the tail pointer position of the voltage ring buffer.

[0095] Step S24) cycle correction step, according to the offset S calculated in the voltage ring buffer min , the cycle of the head pointer in the current ring buffer is corrected, and the formula is

[0096]

[0097] wherein, represents the corrected cycle current at the head pointer of the current ring buffer, I h represents the current at the head pointer of the current ring buffer.

[0098] Step S25) difference operation step, the corrected current cycle at the head pointer of the current ring buffer and the electric cycle at the tail pointer of the current ring buffer are operated, and the differential current cycle is obtained, which is the first cycle, and the formula is

[0099]

[0100] wherein, I diff represents the differential current cycle, I t represents the current at the tail pointer of the current ring buffer.

[0101] Step S3) high frequency feature calculation step, the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude of the three-phase motor are extracted, the characteristic zero rate of the first cycle, the U-I trajectory envelope area and the margin are calculated, the characteristic zero rate, the U-I trajectory envelope area, the margin, the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude are constructed into a high frequency feature set, the first cycle of the three-phase motor which satisfies the high frequency feature of the high frequency feature set is marked, and the second cycle is obtained.

[0102] In this embodiment, the first harmonic amplitude is the 5th harmonic amplitude, the second harmonic amplitude is the 7th harmonic amplitude, and the third harmonic amplitude is the 23rd harmonic amplitude.

[0103] As shown in Figure 3 , step S3) high frequency feature calculation step specifically includes step S31) harmonic amplitude extraction step, step S32) zero rate definition step, step S33) envelope area calculation step and step S34) margin calculation step.

[0104] Step S31) a harmonic amplitude extraction step, extracting a first harmonic amplitude, a second harmonic amplitude and a third harmonic amplitude of the three-phase motor by a discrete Fourier transform formula

[0105]

[0106] wherein f = 0, 1, 2,..., N-1, f represents a frequency generated when the motor operates; N represents the length of a cycle; W represents a complex matrix; n represents an index of the length of a cycle, representing the position in the cycle.

[0107] Step S32) a zero rate self-defining step, depicting the proportion of values near zero in the cycle current by self-defining a zero rate, and the formula is

[0108]

[0109] wherein count low represents the number of values near zero, n represents the number of cycles, n = 32, I i represents the cycle current, lowbound represents the upper threshold of the definition of "values near zero", and minorrate represents the zero rate.

[0110] Step S33) an envelope area calculation step, calculating the envelope area of the relationship curve between the voltage and the current of each phase of the three-phase motor, and the formula is

[0111]

[0112] wherein S U-I represents the envelope area, n represents the number of cycles, n = 32, I i represents the cycle current, U i represents the cycle voltage.

[0113] Step S34) a margin calculation step, calculating the sharpness of the current peak in the waveform, and the formula is

[0114]

[0115] wherein x r represents the square root amplitude of the current, n represents the number of cycles, n = 32, I i represents the cycle current, x peak represents the amplitude of the current, C e represents the calculated margin.

[0116] Further, the judgment standard of whether the high frequency feature set meets the high frequency feature of the three-phase motor is that the first harmonic amplitude is greater than a first amplitude threshold, the second harmonic amplitude is greater than a second amplitude threshold, the third harmonic amplitude is greater than a third amplitude threshold, the zero rate is less than a zero rate threshold, the U-I trajectory envelope area is within a preset area threshold range, and the margin is within a preset margin threshold range.

[0117] In the embodiment, the first amplitude threshold is 1.5, the second amplitude threshold is 0.7, the third amplitude threshold is 0.3, the zero rate threshold is 0.2, the preset area threshold range is 500<S U-I <1200, and the margin threshold range is 1<C e <3.5.

[0118] Step S4) low frequency event detection step, by creating a ring buffer, the size of the ring buffer is 1s, there are 10 positions in the ring buffer for storing active power, reactive power and high frequency feature marks, the active power and the reactive power of each cycle in the ring buffer are calculated, if there is a second cycle in the cycle, the data points of the second cycle are marked with high frequency feature, the data points are 50Hz data points, this mark can help the subsequent processing process to easily identify and process this data point, each position in the ring buffer is used to store the active power, the reactive power and the high frequency feature mark, if the active power and the reactive power stored in a position meet the preset condition, it is determined that there is a low frequency event.

[0119] In the motor operation state monitoring, the detection of low frequency event is crucial to timely discover the abnormal operation of the motor, by analyzing the motor operation data, the active power and the reactive power change in the motor operation and the extreme point of the reactive power are accurately detected, so as to judge whether a specific event occurs and record and store the event.

[0120] As Figure 4 shown, step S4) low frequency event detection step specifically includes step S41) active power calculation step, step S42) reactive power calculation step, step S43) active power accumulation and calculation step and step S44) reactive power accumulation and calculation step.

[0121] Step S41) active power calculation step, the active power of each cycle in the ring buffer is calculated, and the formula is

[0122] P=UIcosφ

[0123] Wherein, P represents the active power, U represents the cycle voltage, and I represents the cycle current.

[0124] Step S42) reactive power calculation step, calculating the reactive power of each cycle in the ring buffer, whose formula is

[0125] Q = U I sinφ

[0126] wherein Q represents the reactive power.

[0127] Step S43) active power cumulative sum calculation step, calculating the cumulative sum of the active power of each cycle in the ring buffer, whose formula is

[0128]

[0129] wherein i represents the time, P k represents the active power at time k, P0 represents the initial value of the active power, S p (i) represents the cumulative sum of the active power.

[0130] Step S44) reactive power cumulative sum calculation step, calculating the cumulative sum of the reactive power of each cycle in the ring buffer, whose formula is

[0131]

[0132] wherein Q k represents the reactive power at time k, Q0 represents the initial value of the reactive power, S Q (i) represents the cumulative sum of the reactive power.

[0133] Further, the preset condition is that if there is a position m satisfying the following two preset conditions, the mth position and the 0th position are determined to constitute a low-frequency event:

[0134]

[0135] wherein 1≤m≤n, abs(S P (m)) represents the cumulative sum of the data points of the active power, abs(S Q (m)) represents the cumulative sum of the data points of the reactive power.

[0136] In this embodiment, when the motor starts, the electromagnetic conversion process will cause significant changes in reactive power. At the moment of motor starting, the stator winding needs to establish a rotating magnetic field, which requires a large amount of reactive current to support, causing the reactive power to rise sharply. When the motor is turned off, the reactive power drops sharply. Based on these significant characteristics of the reactive power during motor starting, the key events in the motor operation can be analyzed by comprehensive judgment of the reactive power and the active power.

[0137] Step S5) An operating status inspection step is performed, wherein a high-frequency feature inspection is performed on the low-frequency events detected in each phase of the three-phase motor. If a high-frequency feature exists, a low-frequency feature inspection is performed on the low-frequency events detected in each phase of the three-phase motor. The operating status of each phase is determined based on the result of the low-frequency feature inspection. The operating status of the three-phase motor is then determined based on the operating status of each phase and the time error when the operating status is turned on.

[0138] like Figure 5 As shown, step S5) the running state inspection step specifically includes step S51) a high frequency feature inspection step, step S52) a low frequency feature inspection step and step S53) a comprehensive determination step.

[0139] Step S51) High-frequency feature detection step: for each low-frequency event detected by phase separation, detect whether the low-frequency event has high-frequency features. The data point set included in the low-frequency event is E = {e1, e2, ..., e n}, where n represents the number of data points contained in the event, e i represents the i-th data point, i=1,2,...,n; define the labeling function M(e i ), used to determine the data point e i Is it marked? When the data point e i When marked, M(e i )=1, and execute the next step; when the data point e i When not marked, M(e i )=0, and detect the next low-frequency event.

[0140] Step S52) Low-frequency feature inspection step, defining the low-frequency event that passes the high-frequency feature inspection step as a marked low-frequency event, and judging whether the marked low-frequency event has a transient low-frequency feature of motor startup. If so, and the reactive power of the marked low-frequency event is within the power threshold range, it is determined that a motor event is detected in the phase splitting.

[0141] Step S53) A comprehensive determination step is performed to determine whether the three-phase motor is in one of the three states of three-phase motor on, three-phase motor missing phase, and three-phase motor off based on the motor events detected in each phase.

[0142] like Figure 6 As shown, step S52) the low frequency feature inspection step specifically includes step S54) a peak detection step, step S55) a peak ratio detection step and step S56) a power jump detection step.

[0143] Step S54) Peak detection step, by detecting the first marker low frequency event E1 = {e 11 ,e 12 ,...,e1n} and its adjacent next second marked low frequency event E2 = {e 21 ,e 22 ,...,e 2n} whether it meets the following detection condition, if it meets, the next step is executed, if it does not meet, the next marked low frequency event pair is detected, the detection condition is

[0144]

[0145] e 21 -e 1n ==1

[0146] Wherein, P represents active power, Q represents reactive power.

[0147] Step S55) peak ratio detection step, judges whether the peak ratio meets the determination condition, if it meets, the next step is executed, if it does not meet, the next peak ratio is detected, the determination condition is

[0148]

[0149] 1.2 < g < 2.1

[0150] Wherein, g represents the peak ratio.

[0151] Step S56) power jump detection step, judges whether the reactive power of the low frequency event is in the power threshold range by merging the active power change of the first marked low frequency event E1 = {e 11 ,e 12 ,...,e 1n} and the second marked low frequency event E2 = {e 21 ,e 22 ,...,e 2n}, the reactive power calculation formula of the low frequency event is

[0152] qdiff = Q[e 2n ]- Q[e 11 ]

[0153] Wherein, qdiff represents the reactive power of the low frequency event, when 50 < qdiff < 300, it is determined that the reactive power of the low frequency event is in the power threshold range, and there is a motor start event.

[0154] In the embodiment, for each phase detected event, further high frequency feature test is performed on the event data points, i.e. determining whether the motor high frequency feature is detected within the event occurrence time period. If the high frequency feature exists, further determination is made on whether the motor start transient low frequency feature exists. The motor start transient will produce a large current change, and then the current sharply drops, which reflects in the power level, and will produce active and reactive power spikes. The low frequency feature section first performs spike detection on the marked low frequency event, if the spike detection passes, further test is performed on the active and reactive power spike peak value variation ratio, if the variation ratio is greater than 1.2 and less than 2.1, it is determined that the motor low frequency power section feature is satisfied, if the above conditions are all satisfied and the event reactive power is greater than 50 var and less than 300 var, it is determined that a motor start event is detected by the phase detection.

[0155] As shown in Figure 7 , the step S53) comprehensive determination step specifically includes a step S57) three-phase motor start determination step, a step S58) three-phase motor open-phase determination step and a step S59) three-phase motor shutdown determination step.

[0156] The step S57) three-phase motor start determination step, when each phase detects a motor start event, the maximum time deviation of the three phase motor start event occurrence times is calculated, and the formula is

[0157] Δt max (t1,t2,t3)=max(|t1-t2|,|t2-t3|,|t1-t3|)

[0158] Wherein, Δt max represents the maximum time deviation, t1, t2, t3 respectively represent the motor start event occurrence time, when Δt max <1, it is determined that the three-phase motor is in the start state.

[0159] In the embodiment, if the three phases all detect the motor start event according to the above steps, at this time, it cannot be simply determined that the three-phase motor is in the start state, because the synchronization of the three-phase motor is very important. Therefore, the maximum time deviation of the three motor event occurrence times is calculated, and the purpose is to consider the coordination of the three-phase power driven motor start.

[0160] The step S58) three-phase motor open-phase determination step, under the premise of detecting the three-phase motor start event, when there is a phase detecting a motor shutdown event within a predetermined time period, and other phases have motors running, it is determined that the three-phase motor is in the open-phase state.

[0161] In the embodiment, on the premise that the "three-phase motor start" event has been clearly verified, the system keeps monitoring the motor running state. Once the motor stop event is detected in a certain phase within 1s, while other phases have motors running, the inconsistent running state indicates the "three-phase motor open-phase" fault condition. The 1s time limit is set based on the characteristics of the rapid changes in electrical parameters and mechanical performance after the motor open-phase, which can capture the open-phase hidden danger at an early stage and avoid further deterioration of the fault. On the other hand, it prevents the three-phase motor from being in a non-strict time alignment state when it is actually turned off, such as the A-phase motor stop event being 0.2s earlier than the B-phase, which leads to the "three-phase motor stop" being misjudged as "three-phase motor open-phase", causing unnecessary impact.

[0162] Step S59) Three-phase motor stop determination step, when no motor start event is detected in each phase, it is determined that the three-phase motor is in a stop state.

[0163] In the embodiment, when no motor start event is detected in each phase, it means that there is no running motor in the three phases, and the three-phase current, voltage and other data sources show that the motor is in a stationary state. At this time, it can be determined that the "three-phase motor is stopped"; this determination is relatively intuitive, but it also depends on the accurate processing of data by each module in the early stage to ensure that there is no misjudgment due to data errors or interference.

[0164] Step S6) Real-time recording step, for the detected three-phase motor, it is matched by a unique number, and the running state of the three-phase motor corresponding to the number is recorded in real time.

[0165] The application provides a non-invasive three-phase motor running state recognition method, which uses a high-precision current and voltage sensor to synchronously sample the current and voltage signals of different phase lines of the three-phase motor at a sampling frequency of not less than 1600Hz. The data collected by the high-precision current and voltage sensor is processed based on a ring buffer data structure to obtain the differential current cycle. The specific harmonics of the current cycle are extracted based on the fast Fourier transform algorithm, and a high-frequency feature set is constructed by combining other statistical features. The events are detected by calculating the cycle active power, reactive power in the ring buffer and judging the extreme points. The events are processed for the three-phase motor data source, the state of the three-phase motor is judged and recorded, so as to realize comprehensive monitoring of the running state of the three-phase motor, and solve the problem that the existing three-phase motor running state recognition method is not comprehensive and accurate enough in monitoring the three-phase motor.

[0166] The above describes in detail the non-intrusive three-phase motor operation state recognition method provided by the application. The principles and implementation manners of the application are described by using specific examples. The above example is only used to help understand the method of the application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the application, the specific implementation manners and application ranges will be changed. In conclusion, the content of the specification should not be understood as a limitation of the application.

Claims

1. A non-invasive method for identifying the operating state of a three-phase electric machine, characterized in that, Specifically comprising the following steps: The signal sampling step synchronously samples the current change and voltage change of different phase lines of the three-phase motor through a high-precision current-voltage sensor, and converts the collected voltage change and current change into voltage signals and current signals; The preprocessing step pre-processes the current signals and voltage signals based on a ring buffer data structure to obtain a first cycle after differentiation; The high-frequency feature calculation step extracts the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude of the three-phase motor, calculates the characteristic zero rate, the U-I trajectory envelope area and the margin of the first cycle, and constructs the characteristic zero rate, the U-I trajectory envelope area, the margin, the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude into a high-frequency feature set. The first cycle that meets the high-frequency features of the three-phase motor is marked to obtain a second cycle; The low-frequency event detection step creates a ring buffer, calculates the active power and reactive power in each cycle in the ring buffer, and if there is a second cycle in the cycle, marks the data points of the second cycle with high-frequency features. Each position in the ring buffer is used to store the active power, the reactive power and the high-frequency feature mark. If the active power and the reactive power stored in a position meet the preset conditions, it is determined that there is a low-frequency event; And The running state inspection step inspects the high-frequency features of the low-frequency events detected by each split phase of the three-phase motor. If there is a high-frequency feature, the low-frequency features of the low-frequency events detected by each split phase of the three-phase motor are inspected. According to the results of the low-frequency feature inspection, the running state of each split phase is determined. The running state of the three-phase motor is determined by the running state of each split phase and the time error of the running state.

2. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The preprocessing step specifically comprises the following steps: The buffer establishment step sets up a voltage ring buffer and a current ring buffer. Each buffer position stores the data of one cycle. The maximum value alignment step finds the maximum value index of the head pointer and the tail pointer by traversal, and aligns the tail pointer voltage with the head pointer voltage. The calculation formula is wherein, represents the voltage at the aligned voltage ring buffer head pointer, k = 0, 1,..., M-1, k represents the data point of the cycle, V h represents the voltage at the voltage ring buffer head pointer position, represents the maximum index of the voltage ring buffer tail pointer, represents the maximum index of the voltage ring buffer head pointer, M = 32; a Manhattan distance calculation step, fixing the cycle at the voltage ring buffer tail pointer, setting an offset, traversing all offsets to calculate the Manhattan distance between the cycle at the voltage ring buffer head pointer and the cycle at the voltage ring buffer tail pointer, recording the minimum distance d min and the corresponding offset s min The calculation formula is where d(s) denotes the Manhattan distance, s denotes the offset, s = -5, -4,..., 4, 5, V t denotes the voltage at the tail pointer position of the voltage ring buffer; a step of correcting the cycle according to the offset S calculated in the voltage ring buffer min a step of correcting the cycle of the head pointer in the current ring buffer, the formula of which is wherein, represents the corrected current ring buffer head pointer at the cycle current, I h represents the current at the head pointer in the current ring buffer; and The difference operation step performs difference operation on the corrected current cycle at the head pointer of the current ring buffer and the current cycle at the tail pointer of the current ring buffer to obtain the differentiated current cycle, which is the first cycle. The formula is wherein I diff represents the differential current cycle, I t represents the current at the tail pointer of the current ring buffer; 3. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The high-frequency feature calculation step specifically comprises the following steps: The harmonic amplitude extraction step extracts the first harmonic amplitude, the second harmonic amplitude and the third harmonic amplitude of the three-phase motor through the discrete Fourier transform formula. The discrete Fourier transform formula is Where f=0,1,2,..., N-1, f represents the frequency generated when the motor is running, N represents the length of the cycle, W represents the complex matrix, n represents the index of the cycle length, and represents the position in the cycle; The zero rate self-definition step defines the zero rate to depict the proportion of the values near the zero point in the cycle current. The formula is wherein, count low represents the number of values around zero, n represents the number of cycles, n = 32, I i represents the cycle current, lowbound represents the upper threshold value satisfying the definition of "values around zero", minorrate represents the zero rate; The envelope area calculation step calculates an envelope area of a curve of relationship between voltage and current of each phase of the three-phase motor, and the formula is where S U-I represents the envelope area, n represents the number of cycles, n = 32, I i represents the cycle current, U i represents the cycle voltage; and The margin calculation step calculates a sharp degree of the current peak in the waveform, and the formula is where x r represents the square root of the amplitude of the current, n represents the number of cycles, n = 32, I i represents the cycle current, x peak represents the amplitude of the current, C e represents the calculated margin.

4. The non-invasive three-phase motor operating condition recognition method of claim 3, wherein, The judgment standard of whether the high frequency feature set meets the high frequency feature of the three-phase motor is that: The first harmonic amplitude is greater than a first amplitude threshold, the second harmonic amplitude is greater than a second amplitude threshold, and the third harmonic amplitude is greater than a third amplitude threshold; The zero rate is less than a zero rate threshold; The U-I trajectory envelope area is within a preset area threshold range; The margin is within a preset margin threshold range.

5. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The low frequency event detection step specifically includes the following steps: The active power calculation step calculates an active power of each cycle in the ring buffer, and the formula is P=UIcosφ wherein P represents the active power, U represents the cycle voltage, and I represents the cycle current; The reactive power calculation step calculates a reactive power of each cycle in the ring buffer, and the formula is Q=UIsinφ wherein Q represents the reactive power; The active power cumulative sum calculation step calculates a cumulative sum of the active power of each cycle in the ring buffer, and the formula is where i represents the time, P k represents the active power at time k, P0represents the initial value of the active power, S p (i) represents the active power cumulative sum; and The reactive power cumulative sum calculation step calculates a cumulative sum of the reactive power of each cycle in the ring buffer, and the formula is where Q k represents the reactive power at time k, Q0represents the initial value of the reactive power, S Q (i) represents the cumulative sum of the reactive power.

6. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The preset condition is that If there is a position m that meets the following two preset conditions, it is determined that the mth position and the 0th position constitute a low frequency event: where 1 < m < n, abs(S P (m)) represents the cumulative sum of data point active power, abs(S Q (m)) represents the cumulative sum of data point reactive power.

7. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The running state inspection step specifically includes the following steps: The high-frequency feature detection step is to detect whether the low-frequency event has high-frequency features for each phase-detected low-frequency event. The data point set included in the low-frequency event is E = {e1, e2, ..., e n }, where n represents the number of data points contained in the event, e i represents the i-th data point, i=1,2,...,n; define the labeling function M(e i ), used to determine the data point e i Is it marked? When the data point e i When marked, M(e i )=1, and execute the next step; when the data point e i When not marked, M(e i )=0, and detect the next low-frequency event; The low frequency feature inspection step defines a low frequency event passing the high frequency feature inspection step as a marked low frequency event, judges whether the marked low frequency event has a motor start transient low frequency feature, and if so, and the reactive power of the marked low frequency event is within a power threshold range, it is determined that a phase detects a motor event; and The comprehensive judgment step judges, based on the motor event detected by each phase, whether the three-phase motor is one of three-phase motor start, three-phase motor open phase and three-phase motor stop.

8. The non-invasive three-phase motor operating condition recognition method of claim 7, wherein, The low frequency feature inspection step specifically includes the following steps: a spike detection step, detecting whether a first marked low frequency event E1 = {e 11 ,e 12 ,...,e 1n} and its adjacent next second marked low frequency event E2 = {e 21 ,e 22 ,...,e 2n} satisfy the following detection condition, if yes, performing the next step, if no, detecting the next marked low frequency event pair, the detection condition is e 21 -e 1n ==1 wherein P represents the active power and Q represents the reactive power; The peak spike ratio detection step judges whether the peak spike ratio meets a judgment condition, and if so, the next step is executed, and if not, the next peak spike ratio is detected, and the judgment condition is that 1.2<g<2.1 wherein g represents the peak spike ratio; and a power jump detection step, merging said first marked low frequency event E1 = {e 11 ,e 12 ,...,e 1n} and said second marked low frequency event E2 = e 21 ,e 22 ,...,e 2n} reactive power variation, determining whether the reactive power of said low frequency event is in a power threshold range, the reactive power of said low frequency event being calculated by the formula qdiff = Q[e 2n ] - Q[e 11 ] wherein qdiff represents the reactive power of the low frequency event, and when 50 9. The non-invasive three-phase motor operating condition recognition method of claim 7, wherein, The comprehensive judgment step specifically includes the following steps: The three-phase motor start judgment step, when each phase detects a motor start event, needs to calculate a maximum time deviation of three-phase motor start event occurrence times, and the formula is Δt max (t1,t2,t3) = max(|t1-t2|, |t2-t3|, |t1-t3|) Wherein, Δt max represents the maximum time deviation, t1, t2, t3 respectively represent the opening event occurrence time of a split-phase motor, when Δt max <1, it is determined that the three-phase motor is in an open state; The three-phase motor open phase judgment step, on the premise of detecting a three-phase motor start event, when there is a phase that detects a motor stop event within a preset time period, and other phases also have a motor in operation, it is determined that the three-phase motor is in an open phase state; and The three-phase motor closing determination step determines that the three-phase motor is in a closed state when no motor starting event is detected in each phase.

10. The non-invasive three-phase motor operating condition recognition method of claim 1, wherein, The method further comprises The real-time recording step records the running state of the three-phase motor corresponding to the number in real time.

Images