A motor fault prediction method for HT80 universal turbine controller

Abstract

A motor fault prediction method for an HT80 universal turbine controller comprises: obtaining pressure signals from the inlet and outlet of a hydraulic servo valve, and performing pulse response decomposition and time-domain envelope analysis on the pressure signals to extract a microbubble nucleation characteristic spectrum from the pressure fluctuations. Based on the microbubble nucleation characteristic spectrum, a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed, and the damping coefficient time series is smoothed using a Kalman filter to obtain a damping coefficient curve. The hydraulic system pulse response model is used to calculate the command response time of the hydraulic servo valve based on the damping coefficient curve, and the command response time is compared with the adjustment target to obtain the response delay. A hysteresis threshold curve is determined through experimental simulation calibration, and based on the hysteresis threshold curve, it is determined whether the response delay exceeds a safety threshold, thereby triggering an alarm. This provides an early warning of the hydraulic response delay trend and can avoid the problem of being unable to complete millisecond-level rapid adjustment during operation.

Description

Technical Field

[0001] The present invention belongs to the technical field of servo motors and fault warning, and more specifically, relates to a motor fault prediction method for an HT80 universal turbine controller. Background Art

[0002] Servo motors are widely used in various industrial fields, and their operating status directly impacts industrial production. A servo motor failure not only disrupts the normal operation of the production line but also increases repair and replacement costs. Therefore, timely and accurate diagnosis of servo motor failures is of great practical significance. Traditional servo motor fault diagnosis methods rely primarily on manual inspection and empirical judgment, which is not only inefficient but also susceptible to human factors.

[0003] In servo motors, viscous hysteresis caused by air intrusion into the servo valve cavity—that is, aging and leakage of the seal between the valve body and cylinder block—draws air into the hydraulic cavity, forming microbubbles. This causes a hydraulic response lag (≥15 ms), making it difficult for the turbine controller to achieve rapid adjustments in the 1ms range. Therefore, it is necessary to provide a prediction method for bubble-induced viscous hysteresis failures in the hydraulic servo valve cavity of the HT80 universal turbine controller. By online monitoring and modeling the microbubble formation mechanism and dynamic response characteristics within the valve cavity, this method can provide early warning of hydraulic response delay trends, thus preventing the inability to achieve rapid adjustments in the millisecond range during operation. Summary of the Invention

[0004] In order to address the deficiencies in the prior art, the present invention aims to solve the above-mentioned defects and further propose a motor fault prediction method for the HT80 universal turbine controller.

[0005] The present invention adopts the following technical solutions.

[0006] A first aspect of the present invention discloses a motor fault prediction method for an HT80 universal turbine controller, the method comprising:

[0007] Obtaining pressure signals at the inlet and outlet of the hydraulic servo valve, and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract microbubble nucleation characteristic spectra from pressure fluctuations;

[0008] Based on the microbubble nucleation characteristic spectrum, a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed, and the damping coefficient time series is smoothed by Kalman filtering to obtain a damping coefficient curve;

[0009] Invoking a hydraulic system pulse response model to calculate a command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with a regulation target to obtain a response delay;

[0010] A hysteresis threshold curve is determined through experimental simulation calibration, and based on the hysteresis threshold curve, it is determined whether the response delay exceeds a safety threshold to trigger an alarm.

[0011] Furthermore, the step of obtaining the pressure signals at the inlet and outlet of the hydraulic servo valve and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract the microbubble nucleation characteristic spectrum in the pressure fluctuation includes:

[0012] Setting pressure sensors at the inlet and outlet of the hydraulic servo valve, and collecting pressure signals of the inlet and outlet of the hydraulic servo valve by the pressure sensors at a set sampling frequency to obtain a sampling sequence;

[0013] The sampling sequence is transformed into an analytical signal by Hilbert transform, and an envelope is calculated based on the analytical signal, so as to apply low-pass filtering to the envelope to obtain a signal envelope.

[0014] Furthermore, the method of obtaining pressure signals at the inlet and outlet of the hydraulic servo valve and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract microbubble nucleation characteristic spectra in the pressure fluctuations further includes:

[0015] Performing Fourier transform on the envelope signal to obtain a discrete spectrum, and extracting frequency domain peaks and broadband features within a typical bubble nucleation frequency band;

[0016] Obtaining a frequency domain peak range and a broadband range through a calibration experiment, and calculating a bubble characteristic index based on the frequency domain peak range and the broadband range;

[0017] Among them, the bubble characteristic index is obtained by normalizing the frequency domain peak and broadband characteristics based on the maximum and minimum values ​​of the frequency domain peak range and broadband range of the inlet and outlet of the hydraulic servo valve, and is divided into the inlet peak index, inlet broadband index, outlet peak index and outlet broadband index of the hydraulic servo valve.

[0018] Furthermore, based on the microbubble nucleation characteristic spectrum, a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed, and the damping coefficient time series is smoothed by Kalman filtering to obtain a damping coefficient curve, including:

[0019] The inlet baseline damping and outlet baseline damping in the bubble-free state are set through static tests, and the peak weight coefficients and broadband weight coefficients of the inlet and outlet cavities are determined;

[0020] According to the peak weight coefficient and broadband weight coefficient of the inlet cavity and the outlet cavity, combined with the inlet peak index, the inlet broadband index, the inlet baseline damping and the outlet peak index, the outlet broadband index, the outlet baseline damping, the equivalent damping coefficient of the inlet cavity and the outlet cavity of the hydraulic servo valve are calculated respectively.

[0021] Furthermore, the mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed based on the microbubble nucleation characteristic spectrum, and the damping coefficient time series is smoothed by Kalman filtering to obtain a damping coefficient curve, which also includes:

[0022] Based on the equivalent damping coefficients of the inlet cavity and the outlet cavity, the maximum value of the equivalent damping coefficients of the inlet cavity and the outlet cavity is selected as the system bottleneck damping at regular intervals to construct an equivalent damping sequence;

[0023] The equivalent damping sequence is filtered by a Kalman filter, and a damping growth rate is calculated based on the fixed time interval to construct the damping coefficient curve.

[0024] Furthermore, the calling of the hydraulic system impulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with the adjustment target to obtain the response delay includes:

[0025] Obtaining the piston equivalent mass of the hydraulic servo valve, and converting the damping coefficient into a real-time response delay based on the damping coefficient time series after Kalman filtering using a first-order system time constant model;

[0026] The response delays of adjacent cycles are determined based on the real-time response delay, and the response delay change rate is estimated according to the response delays of the adjacent cycles using a difference method.

[0027] Furthermore, the calling of the hydraulic system impulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with the adjustment target to obtain the response delay also includes:

[0028] Obtaining the current response delay and the current system damping, combining the response delay change rate and the damping coefficient curve, calculating the physical derivative approximation result and the differential rate statistical result, and performing weighted averaging on the physical derivative approximation result and the differential rate statistical result to obtain a fusion trend slope;

[0029] A linear extrapolation formula is called to calculate an extrapolation result based on the current response delay and the fusion trend slope, and the extrapolation result is stored in a prediction buffer to obtain a response delay prediction sequence.

[0030] Furthermore, the step of determining a hysteresis threshold curve through experimental simulation calibration, and judging whether the response delay exceeds a safety threshold based on the hysteresis threshold curve to trigger an alarm, includes:

[0031] Obtaining a maximum allowable system delay time and a predicted number of steps parameter, and setting a safety threshold according to the maximum allowable system delay time, so as to write the predicted number of steps parameter and the safety threshold into a controller configuration area;

[0032] Based on the current response delay and a safety threshold, determining whether the current response delay exceeds the safety threshold, and marking the current response delay as a first value if the current response delay exceeds the safety threshold, otherwise marking the current response delay as a second value;

[0033] Based on the response delay prediction sequence and a safety threshold, the response delay prediction sequence is scanned to determine the number of times the same response delay is marked as a first value, and a graded alarm is triggered according to the set number of times.

[0034] A second aspect of the present invention discloses a motor fault prediction device for an HT80 universal turbine controller, which is used to implement the motor fault prediction method for an HT80 universal turbine controller described in any one of the first aspects. The device comprises:

[0035] A feature extraction module is used to obtain the pressure signals of the hydraulic servo valve inlet and outlet, and perform pulse response decomposition and time domain envelope analysis on the pressure signals to extract the microbubble nucleation characteristic spectrum in the pressure fluctuation;

[0036] a damping coefficient calculation module, configured to construct a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient based on the microbubble nucleation characteristic spectrum, and smooth the damping coefficient time series through Kalman filtering to obtain a damping coefficient curve;

[0037] a response delay calculation module, configured to call a hydraulic system pulse response model to calculate a command response time of the hydraulic servo valve according to the damping coefficient curve, and compare the command response time with a regulation target to obtain a response delay;

[0038] The alarm trigger module is used to determine a hysteresis threshold curve through experimental simulation calibration, and judge whether the response delay exceeds a safety threshold based on the hysteresis threshold curve to trigger an alarm.

[0039] A third aspect of the present invention discloses a terminal, comprising a processor and a storage medium;

[0040] The storage medium is used to store instructions;

[0041] The processor is configured to operate according to the instructions to execute the steps of the method of the first aspect.

[0042] A fourth aspect of the present invention discloses a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method described in the first aspect when executed by a processor.

[0043] The beneficial effects of the present invention are that, compared with the prior art, the present invention has the following advantages:

[0044] (1) The present invention performs high-frequency sampling on the inlet and outlet pressure and flow signals of the hydraulic servo valve, applies pulse response decomposition and time-domain envelope analysis, and then extracts the microbubble nucleation characteristic spectrum contained in the pressure fluctuation, fully reflecting the instantaneous bubble jumping phenomenon and bubble aggregation range, and providing a basis for judging the change of bubble concentration.

[0045] (2) Based on the obtained bubble characteristic index and combined with the hydraulic system damping theory, this paper constructs a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient. The damping coefficient time series is smoothed by Kalman filtering, revealing how the change in bubble volume affects the system viscosity characteristics and providing parameters for subsequent response delay estimation.

[0046] (3) Based on the damping coefficient curve and the hydraulic system pulse response model, the present invention calculates the actual response time required for the current hydraulic valve to reach its position after the command is issued, and compares it with the millisecond-level rapid adjustment target. Combined with the bubble nucleation and aggregation behavior in fluid dynamics and the dynamic response theory of the hydraulic system, the present invention focuses on the coupling of the bubble generation-growth-aggregation process in the valve cavity and the system's viscous hysteresis characteristics, and establishes a "bubble induction-damping change-response delay" dynamic model. This not only realizes early warning of the hydraulic response delay trend, but also avoids the problem of being unable to complete millisecond-level rapid adjustment during motor operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of a motor fault prediction method for an HT80 universal turbine controller provided by the present invention;

[0048] Figure 2 The present invention provides a schematic structural diagram of a motor fault prediction device for an HT80 universal turbine controller. DETAILED DESCRIPTION

[0049] The present application will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present application.

[0050] like Figure 1 As shown, in one embodiment, a motor fault prediction method for an HT80 universal turbine controller includes the following steps:

[0051] Step S110 , obtaining pressure signals at the inlet and outlet of the hydraulic servo valve, and performing impulse response decomposition and time domain envelope analysis on the pressure signals to extract microbubble nucleation characteristic spectra from the pressure fluctuations.

[0052] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S110 specifically includes the following steps:

[0053] Step S111 : setting pressure sensors at the inlet and outlet of the hydraulic servo valve, and collecting pressure signals of the inlet and outlet of the hydraulic servo valve through the pressure sensors at a set sampling frequency to obtain a sampling sequence.

[0054] Step S112 : transforming the sampling sequence into an analytical signal through Hilbert transform, and calculating an envelope based on the analytical signal to apply low-pass filtering to the envelope to obtain a signal envelope.

[0055] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S110 specifically further includes the following steps:

[0056] Step S113 , performing Fourier transform on the envelope signal to obtain a discrete spectrum, and extracting frequency domain peaks and broadband features within a typical bubble nucleation frequency band.

[0057] Step S114 , obtaining the frequency domain peak range and broadband range through a calibration experiment, and calculating the bubble characteristic index based on the frequency domain peak range and broadband range.

[0058] Among them, the bubble characteristic index is obtained by normalizing the frequency domain peak and broadband characteristics based on the maximum and minimum values ​​of the frequency domain peak range and broadband range of the inlet and outlet of the hydraulic servo valve, and is divided into the inlet peak index, inlet broadband index, outlet peak index and outlet broadband index of the hydraulic servo valve.

[0059] In a specific embodiment, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, comprising steps 1 to 4:

[0060] Step 1: Online feature extraction of valve cavity microbubbles.

[0061] High-frequency sampling of the inlet and outlet pressure and flow signals of the hydraulic servo valve is performed, and impulse response decomposition and time-domain envelope analysis are applied to extract the microbubble nucleation characteristic spectrum contained in the pressure fluctuation.

[0062] The following sub-steps are included:

[0063] Sub-step 1.1, data acquisition.

[0064] Specifically, a pair of high-frequency pressure sensors (measuring range 0-10MPa, broadband 0-10kHz) are installed at the inlet and outlet of the hydraulic servo valve. The inlet and outlet pressure signals output by the pressure sensors are sampled through a 16-bit data acquisition card (DAQ) with a sampling frequency of 20kHz to obtain a time-series sampling sequence, laying the foundation for subsequent envelope and spectrum analysis.

[0065] Sub-step 1.2: Time domain envelope extraction.

[0066] Specifically, first, the sampling sequence obtained in sub-step 1.1 Perform Hilbert transform to obtain the corresponding analytical signal sequence :

[0067]

[0068] Where, is the Hilbert coefficient, is the Hilbert transform operator.

[0069] Afterwards, calculate the envelope :

[0070]

[0071] Finally, the envelope Apply 1kHz low-pass filtering to remove subharmonic interference and retain the bubble nucleation frequency band (0-1kHz), where the envelope It can highlight the high-frequency oscillation components of microbubbles and facilitate quantification of their characteristics.

[0072] It should be noted that the sampling sequence Divided into two groups, namely import sampling sequence and export sampling sequence Similarly, the signal sequence is analyzed according to the above calculation formula and envelope Also divided into two groups.

[0073] Sub-step 1.3: Frequency domain feature analysis.

[0074] Specifically, the envelope calculated in step 1.2 Perform fast Fourier transform (FFT) to obtain discrete spectrum , where k=1,2,..., , frequency resolution , is the sampling frequency, is the number of FFT points. Within the typical bubble nucleation frequency range (100-500 Hz), the maximum amplitude (i.e., the frequency domain peak, including both the outlet and inlet frequency domain peaks) and the 3dB bandwidth are extracted. The frequency domain peak reflects the instantaneous bubble jump phenomenon, while the bandwidth reflects the bubble aggregation range, providing a basis for determining changes in bubble concentration.

[0075] Sub-step 1.4, characteristic index calculation and normalization.

[0076] Specifically, based on the frequency domain peak and 3dB bandwidth obtained in sub-step 1.3, the peak range is obtained according to the calibration test and broadband range , the bubble characteristic index of the servo valve inlet and outlet is calculated by normalization:

[0077]

[0078]

[0079]

[0080]

[0081] Where, They are import peak index, export peak index, import broadband index and export broadband index. They represent the 3dB bandwidth corresponding to the maximum point of the servo valve inlet pressure amplitude and the 3dB bandwidth corresponding to the maximum point of the servo valve outlet pressure amplitude, respectively. and They represent the points with the maximum pressure amplitude at the servo valve inlet and the maximum pressure amplitude at the servo valve outlet respectively.

[0082] It should be noted that the calibration experiment In the laboratory or on-site, first build a test platform with the same valve cavity geometry and oil system as the actual valve. Then, obtain the envelope peak value and the upper and lower limits of the bandwidth according to the following process:

[0083] (1) Clean baseline measurement: Replace the new, vacuum-degassed hydraulic oil to ensure that there are almost no microbubbles in the valve cavity; perform sub-steps 1.1 to 1.3 on it, and the maximum amplitude point and 3dB bandwidth collected are respectively used as .

[0084] (2) Controlled bubble injection: In the oil circuit, a precision microinjection pump is used to inject air microbubbles into the oil at volume fractions of 0.1%, 0.2%, 0.5%, and 1.0%. To simulate extreme failure scenarios, the bubble concentration is increased to 2%–5% until the envelope oscillation signal becomes saturated or masked by noise. Then, substeps 1.1 to 1.3 are executed, and the maximum amplitude point and 3dB bandwidth collected are used as the .

[0085] Step S120 : constructing a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient based on the microbubble nucleation characteristic spectrum, and smoothing the damping coefficient time series through Kalman filtering to obtain a damping coefficient curve.

[0086] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S120 specifically includes the following steps:

[0087] Step S121 , setting the inlet baseline damping and the outlet baseline damping in a bubble-free state through a static test, and determining the peak weight coefficients and broadband weight coefficients of the inlet cavity and the outlet cavity.

[0088] Step S122, based on the peak weight coefficients and broadband weight coefficients of the inlet and outlet cavities, combined with the inlet peak index, inlet broadband index, inlet baseline damping and the outlet peak index, outlet broadband index, outlet baseline damping, respectively calculate the equivalent damping coefficients of the inlet and outlet cavities of the hydraulic servo valve.

[0089] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S120 specifically further includes the following steps:

[0090] Step S123 , based on the equivalent damping coefficients of the inlet cavity and the outlet cavity, the maximum value is selected from the equivalent damping coefficients of the inlet cavity and the outlet cavity as the system bottleneck damping at regular intervals to construct an equivalent damping sequence.

[0091] Step S124 : filtering the equivalent damping sequence by using a Kalman filter, and calculating the damping growth rate based on a fixed time interval to construct a damping coefficient curve.

[0092] In a specific embodiment, the present invention provides a motor fault prediction method for an HT80 universal turbine controller. Step 2 involves modeling bubble aggregation and damping variation. Based on the obtained bubble characteristic index and combined with hydraulic system damping theory, a mapping relationship between the bubble volume fraction and the system's equivalent damping coefficient is constructed. A Kalman filter is used to smooth the damping coefficient time series, revealing how changes in bubble volume affect the system's viscosity characteristics and providing parameters for response delay estimation.

[0093] The following sub-steps are included:

[0094] Sub-step 2.1: Calculation of local damping in the inlet cavity.

[0095] Specifically, based on the import peak index obtained in step 1 and Import Broadband Index , through static experiments to set the inlet baseline damping in the bubble-free state, and determine the peak weight coefficient and broadband weight coefficient at the same time, and then calculate the equivalent damping coefficient of the inlet cavity :

[0096]

[0097] Where, , is the inlet baseline damping in the bubble-free state; , is the peak weight coefficient of the inlet cavity; , is the broadband weight coefficient of the inlet cavity.

[0098] It should be noted that the peak weight and bandwidth weight During the acquisition process, fixed ,Change , record the corresponding damping increment to determine the degree of influence of peak change on damping; then, conversely, fix ,Change , record the corresponding damping increment, and then determine the degree to which the bandwidth change affects the damping, thereby determining the peak weight and bandwidth weight respectively. The peak weight coefficient and broadband weight coefficient of the inlet cavity can be obtained similarly.

[0099] Sub-step 2.2: Calculation of local damping of the outlet cavity.

[0100] Specifically, based on the export peak index obtained in step 1 and export broadband index , through static experiments to set the outlet baseline damping in the bubble-free state, and at the same time determine the peak weight coefficient and broadband weight coefficient, and then calculate the oral equivalent damping coefficient :

[0101]

[0102] Where, , is the outlet baseline damping in the bubble-free state; , is the peak weight coefficient of the outlet cavity; , is the broadband weight coefficient of the outlet cavity.

[0103] Sub-step 2.3: Determine the equivalent damping of the entire system.

[0104] Specifically, based on the equivalent damping coefficient of the inlet cavity obtained in sub-steps 2.1 and 2.2 and the equivalent damping coefficient of the outlet cavity , take the larger value of the two damping as the system bottleneck damping :

[0105]

[0106] This is because the cavity where bubble aggregation is most serious determines the maximum damping of the overall hydraulic pressure. Other lower damping paths are difficult to make up for the bottleneck effect. The maximum damping highlights the damping characteristics of the most unfavorable path, which can prevent the low damping channel from masking the delay risk brought by the high damping channel.

[0107] It should be noted that sub-steps 2.1 to 2.3 are continuously calculated at a fixed interval of , so the output of sub-step 2.3 is actually .

[0108] Sub-step 2.4, filtering and trend monitoring.

[0109] Specifically, based on the system bottleneck damping determined in sub-step 2.3 , construct the equivalent damping sequence , and perform Kalman filtering on each damping in the equivalent damping sequence:

[0110]

[0111] Where, For sequence The damping after the k-th damping filter, is the filter gain.

[0112] Afterwards, the damping growth rate is calculated based on the Kalman filter results :

[0113]

[0114] It should be noted that Refers to the "post-filter damping" value calculated at the last sampling moment of k (i.e., the previous cycle), which can be agreed to be .

[0115] Step S130 , calling the hydraulic system pulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with the adjustment target to obtain the response delay.

[0116] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S130 specifically includes the following steps:

[0117] Step S131 : obtaining the piston equivalent mass of the hydraulic servo valve, and converting the damping coefficient into a real-time response delay based on the damping coefficient time series after Kalman filtering through a first-order system time constant model.

[0118] Step S132 : determining the response delay of adjacent cycles based on the real-time response delay, and estimating the response delay change rate according to the response delay of adjacent cycles using a difference method.

[0119] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S130 specifically further includes the following steps:

[0120] Step S133: Obtain the current response delay and the current system damping, combine the response delay change rate and the damping coefficient curve, calculate the physical derivative approximation result and the differential rate statistical result, and perform weighted average on the physical derivative approximation result and the differential rate statistical result to obtain the fusion trend slope.

[0121] Step S134 , calling a linear extrapolation formula to calculate an extrapolation result based on the current response delay and the fusion trend slope, and storing the extrapolation result in a prediction buffer to obtain a response delay prediction sequence.

[0122] In a specific embodiment, the present invention provides a motor fault prediction method for an HT80 universal turbine controller. Step 3, response delay trend estimation, calculates the actual response time required for the hydraulic valve to reach its desired position after a command is issued based on the damping coefficient curve and the hydraulic system impulse response model, and compares this with the millisecond-level rapid adjustment target.

[0123] The following sub-steps are included:

[0124] Sub-step 3.1, real-time response delay calculation.

[0125] Specifically, based on the filtered equivalent damping obtained in step 2 (N·s / m), and at the same time obtain the equivalent mass of the piston m (kg). According to the first-order system time constant model, the equivalent damping after filtering is Converting to real-time response latency :

[0126]

[0127] It should be noted that the piston is located in the hydraulic servo valve assembly. As you can imagine, pure, bubble-free hydraulic oil is virtually incompressible, allowing oil circuit pressure to act directly on the piston and be transmitted rapidly. However, once bubbles are introduced into the oil, they compress like a spring, absorbing some of the pressure energy, resulting in slower piston displacement under the same pressure fluctuations. Bubbles aggregate and rupture within the valve cavity, dramatically altering local damping in a short period of time. Piston movement relies on oil film friction and oil circuit damping. If the damping suddenly increases or fluctuates, the piston response time will be prolonged or jitter will occur, resulting in a delay.

[0128] Sub-step 3.2, delay change rate calculation.

[0129] Specifically, first determine the current response delay and the previous cycle response delay , using the difference method to estimate the instantaneous rate :

[0130]

[0131] The instantaneous rate is then cached for trend slope calculation, which explains whether the response latency is increasing or decreasing, providing first-order information for short-term prediction.

[0132] Sub-step 3.3, trend slope fusion.

[0133] Specifically, the instantaneous rate calculated based on the above steps , damping growth rate , Current response delay and the current damping , computational physics derivative approximation :

[0134]

[0135] and statistical differential rate , and then perform a weighted average of the two, with each weight being 0.5, to calculate the fusion trend slope :

[0136]

[0137] Sub-step 3.3 combines statistical and physical models to accurately capture latency variation trends and improve prediction robustness.

[0138] Sub-step 3.4, short-term delay prediction.

[0139] Specifically, based on the current response delay obtained in the previous steps and fusion slope trend , perform linear extrapolation:

[0140]

[0141] Where, To predict the number of steps; To predict the total number of steps, .

[0142] Afterwards, the linear extrapolation results are stored in the prediction buffer for use in alarms. This step provides a prediction of the response delay change within 0.25 seconds based on the current trend, making it easier to take motor control or maintenance measures in advance.

[0143] Step S140 , determining a hysteresis threshold curve through experimental simulation calibration, and judging whether the response delay exceeds a safety threshold based on the hysteresis threshold curve to trigger an alarm.

[0144] In some embodiments, the present invention provides a motor fault prediction method for an HT80 universal turbine controller, wherein step S140 specifically includes the following steps:

[0145] Step S141 , obtaining the system's maximum allowable delay time and predicted step number parameters, and setting a safety threshold according to the system's maximum allowable delay time, so as to write the predicted step number parameters and the safety threshold into the controller configuration area.

[0146] Step S142: Based on the current response delay and the safety threshold, determine whether the current response delay exceeds the safety threshold, and mark the current response delay as a first value if the current response delay exceeds the safety threshold, otherwise mark it as a second value.

[0147] Step S143 : Based on the response delay prediction sequence and the safety threshold, the response delay prediction sequence is scanned to determine the number of times the same response delay is marked as the first value, and a graded alarm is triggered according to the set number of times.

[0148] In a specific embodiment, the present invention provides a motor fault prediction method for an HT80 universal turbine controller. Step 4, hysteresis anomaly detection and threshold alarm, determines whether the current response delay exceeds a safety warning line based on a hysteresis threshold curve obtained through experimental or simulation calibration. Repeated over-limit exceeding the limit triggers a high-level alarm.

[0149] The following sub-steps are included:

[0150] Sub-step 4.1: Delay threshold and alarm level setting.

[0151] Specifically, first, a safety threshold of 3s is set according to the maximum period allowed for rapid adjustment on site, and then the safety threshold is written into the controller configuration area.

[0152] Sub-step 4.2: Real-time delay overrun detection.

[0153] Specifically, based on the current response delay and the safety threshold set in sub-step 4.1 , make comparison and judgment, and get real-time over-limit mark :

[0154]

[0155] If the over-limit flag is equal to 1, the current delay value and timestamp are recorded in the register, otherwise the record is cleared to immediately capture events where the response delay exceeds 3s and trigger a high-priority alarm judgment.

[0156] Sub-step 4.3, predict delay overrun detection.

[0157] Specifically, based on the future short-term delay predictions and safety thresholds obtained in the previous steps, if any predicted value in the future short-term delay predictions exceeds the safety threshold, an overlimit flag is set to 1; otherwise, it is set to 0. When the overlimit flag is 1, the predicted step and its corresponding predicted value of the first overlimit are recorded to preemptively identify the risk of overlimit within the next 0.25 seconds, facilitating early intervention in motor maintenance and control strategies.

[0158] Sub-step 4.4: continuous over-limit counting and graded alarm.

[0159] Specifically, based on the real-time over-limit flag, predicted over-limit flag, and continuous count obtained in the above steps, any over-limit flag and its maintenance counter are determined, and a graded alarm is issued:

[0160] If the real-time over-limit flag value is 1 and the maintenance counter is ≥ 2, a "serious alarm" is output;

[0161] If the real-time over-limit flag value is 1, then the "real-time alarm" is output;

[0162] If the predicted over-limit flag value is 1, the "prediction alarm" is output;

[0163] If all of the above situations occur, the output is "no alarm".

[0164] During the alarm output process, the DO / I / O module drives the physical relay indication, reports to the HMI / SCADA interface and records the alarm log. Continuous counting can suppress instantaneous jitter false alarms, and combines real-time and predictive information for graded alarms, ensuring timely and accurate response to delay anomalies.

[0165] The following describes a motor fault prediction device for an HT80 universal turbine controller provided by the present invention. The motor fault prediction device for an HT80 universal turbine controller described below and the motor fault prediction method for an HT80 universal turbine controller described above can be referenced to each other.

[0166] like Figure 2 As shown, in one embodiment, a motor fault prediction device for an HT80 universal turbine controller includes a feature extraction module, a damping coefficient calculation module, a response delay calculation module, and an alarm triggering module.

[0167] The feature extraction module is used to obtain the pressure signals of the hydraulic servo valve inlet and outlet, and perform pulse response decomposition and time domain envelope analysis on the pressure signals to extract the microbubble nucleation characteristic spectrum in the pressure fluctuation.

[0168] The damping coefficient calculation module is used to construct a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient based on the microbubble nucleation characteristic spectrum, and smooth the damping coefficient time series through Kalman filtering to obtain the damping coefficient curve.

[0169] The response delay calculation module is used to call the hydraulic system pulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and compare the command response time with the adjustment target to obtain the response delay.

[0170] The alarm trigger module is used to determine the hysteresis threshold curve through experimental simulation calibration, and judge whether the response delay exceeds the safety threshold based on the hysteresis threshold curve to trigger an alarm.

[0171] The applicant of the present invention has made a detailed explanation and description of the implementation examples of the present invention in conjunction with the drawings in the specification. However, those skilled in the art should understand that the above implementation examples are only preferred implementation plans of the present invention, and the detailed description is only to help readers better understand the spirit of the present invention, and is not a limitation on the scope of protection of the present invention. On the contrary, any improvements or modifications based on the inventive spirit of the present invention should fall within the scope of protection of the present invention.

Claims

1. A motor fault prediction method for HT80 universal turbine controller, characterized in that: The method comprises: Obtaining pressure signals at the inlet and outlet of the hydraulic servo valve, and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract microbubble nucleation characteristic spectra from pressure fluctuations; Based on the microbubble nucleation characteristic spectrum, a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed, and the damping coefficient time series is smoothed by Kalman filtering to obtain a damping coefficient curve; Invoking a hydraulic system pulse response model to calculate a command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with a regulation target to obtain a response delay; A hysteresis threshold curve is determined through experimental simulation calibration, and based on the hysteresis threshold curve, it is determined whether the response delay exceeds a safety threshold to trigger an alarm.

2. The motor fault prediction method for the HT80 universal turbine controller according to claim 1, characterized in that: The step of obtaining the pressure signals of the hydraulic servo valve inlet and outlet, and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract the microbubble nucleation characteristic spectrum in the pressure fluctuations includes: Setting pressure sensors at the inlet and outlet of the hydraulic servo valve, and collecting pressure signals of the inlet and outlet of the hydraulic servo valve by the pressure sensors at a set sampling frequency to obtain a sampling sequence; The sampling sequence is transformed into an analytical signal by Hilbert transform, and an envelope is calculated based on the analytical signal, so as to apply low-pass filtering to the envelope to obtain a signal envelope.

3. The motor fault prediction method for the HT80 universal turbine controller according to claim 2, characterized in that: The method of obtaining pressure signals at the inlet and outlet of the hydraulic servo valve and performing pulse response decomposition and time domain envelope analysis on the pressure signals to extract microbubble nucleation characteristic spectra in the pressure fluctuations further includes: Performing Fourier transform on the envelope signal to obtain a discrete spectrum, and extracting frequency domain peaks and broadband features within a typical bubble nucleation frequency band; Obtaining a frequency domain peak range and a broadband range through a calibration experiment, and calculating a bubble characteristic index based on the frequency domain peak range and the broadband range; Among them, the bubble characteristic index is obtained by normalizing the frequency domain peak and broadband characteristics based on the maximum and minimum values ​​of the frequency domain peak range and broadband range of the inlet and outlet of the hydraulic servo valve, and is divided into the inlet peak index, inlet broadband index, outlet peak index and outlet broadband index of the hydraulic servo valve.

4. The motor fault prediction method for the HT80 universal turbine controller according to claim 3, characterized in that: The mapping relationship between the bubble volume fraction and the system equivalent damping coefficient is constructed based on the microbubble nucleation characteristic spectrum, and the damping coefficient time series is smoothed by Kalman filtering to obtain a damping coefficient curve, including: The inlet baseline damping and outlet baseline damping in the bubble-free state are set through static tests, and the peak weight coefficients and broadband weight coefficients of the inlet and outlet cavities are determined; According to the peak weight coefficient and broadband weight coefficient of the inlet cavity and the outlet cavity, combined with the inlet peak index, the inlet broadband index, the inlet baseline damping and the outlet peak index, the outlet broadband index, the outlet baseline damping, the equivalent damping coefficient of the inlet cavity and the outlet cavity of the hydraulic servo valve are calculated respectively.

5. The motor fault prediction method for the HT80 universal turbine controller according to claim 4, characterized in that: The method further includes: constructing a mapping relationship between the bubble volume fraction and the system equivalent damping coefficient based on the microbubble nucleation characteristic spectrum, and smoothing the damping coefficient time series through Kalman filtering to obtain a damping coefficient curve. Based on the equivalent damping coefficients of the inlet cavity and the outlet cavity, the maximum value of the equivalent damping coefficients of the inlet cavity and the outlet cavity is selected as the system bottleneck damping at regular intervals to construct an equivalent damping sequence; The equivalent damping sequence is filtered by a Kalman filter, and a damping growth rate is calculated based on the fixed time interval to construct the damping coefficient curve.

6. The motor fault prediction method for the HT80 universal turbine controller according to claim 1, characterized in that: The calling of the hydraulic system impulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with the adjustment target to obtain the response delay includes: Obtaining the piston equivalent mass of the hydraulic servo valve, and converting the damping coefficient into a real-time response delay based on the damping coefficient time series after Kalman filtering using a first-order system time constant model; The response delays of adjacent cycles are determined based on the real-time response delay, and the response delay change rate is estimated according to the response delays of the adjacent cycles using a difference method.

7. The motor fault prediction method for the HT80 universal turbine controller according to claim 6, characterized in that: The calling of the hydraulic system impulse response model to calculate the command response time of the hydraulic servo valve according to the damping coefficient curve, and comparing the command response time with the adjustment target to obtain the response delay further includes: Obtaining the current response delay and the current system damping, combining the response delay change rate and the damping coefficient curve, calculating the physical derivative approximation result and the differential rate statistical result, and performing weighted averaging on the physical derivative approximation result and the differential rate statistical result to obtain a fusion trend slope; A linear extrapolation formula is called to calculate an extrapolation result based on the current response delay and the fusion trend slope, and the extrapolation result is stored in a prediction buffer to obtain a response delay prediction sequence.

8. The motor fault prediction method for the HT80 universal turbine controller according to claim 7, characterized in that: The step of determining a hysteresis threshold curve through experimental simulation calibration, and judging whether the response delay exceeds a safety threshold based on the hysteresis threshold curve to trigger an alarm, includes: Obtaining a maximum allowable system delay time and a predicted number of steps parameter, and setting a safety threshold according to the maximum allowable system delay time, so as to write the predicted number of steps parameter and the safety threshold into a controller configuration area; Based on the current response delay and a safety threshold, determining whether the current response delay exceeds the safety threshold, and marking the current response delay as a first value if the current response delay exceeds the safety threshold, otherwise marking the current response delay as a second value; Based on the response delay prediction sequence and a safety threshold, the response delay prediction sequence is scanned to determine the number of times the same response delay is marked as a first value, and a graded alarm is triggered according to the set number of times.

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images