Flywheel life prediction method, system, device, storage medium and program

Abstract

The application discloses a flywheel life prediction method, system, device, storage medium and program, relates to the technical field of flywheel life prediction, and comprises the following steps: collecting an acoustic emission time sequence of a flywheel during operation; calculating a fractal dimension of an acoustic emission signal based on the acoustic emission time sequence; calculating a Lyapunov exponent based on the acoustic emission time sequence; calculating a health index based on the fractal dimension and the Lyapunov exponent; and predicting a remaining life based on the health index. Through the processing scheme, a key time window is provided for predictive maintenance of the flywheel system, so that sudden failures can be avoided, and the reliability and availability of the flywheel system are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of flywheel life prediction technology, and in particular to a method, system, device, storage medium and program for predicting flywheel life. Background Technology

[0002] Failure of flywheel systems often begins with the accumulation of microscopic damage within the material, such as the initiation of microcracks in bearing raceways or the microfracture of fibers in composite flywheels. These microscopic activities release stress waves, i.e., acoustic emission signals. Traditional vibration analysis is not sensitive to early microscopic damage, and conventional acoustic emission techniques rely heavily on simple parameters (such as event counts and energy), which cannot effectively characterize the nonlinear dynamic process of damage evolution, resulting in insufficient early warning capabilities and biased life prediction.

[0003] Therefore, it is evident that the existing flywheel life prediction methods described above still have inconveniences and shortcomings in their use, and urgently need further improvement. Creating a new flywheel life prediction method has become a pressing goal for the industry. Summary of the Invention

[0004] In view of this, the present disclosure provides a method for predicting the lifespan of a flywheel, which at least partially solves the problems existing in the prior art.

[0005] In a first aspect, embodiments of this disclosure provide a method for predicting the lifespan of a flywheel, the method comprising the following steps: Acoustic emission time series of the flywheel during operation; The fractal dimension of the acoustic emission signal is calculated based on the acoustic emission time series. Calculate the Lyapunov exponent based on the acoustic emission time series; The health index is calculated based on the fractal dimension and the Lyapunov index. Predict remaining lifespan based on the aforementioned health index.

[0006] According to a specific implementation of this disclosure, the acquisition of the acoustic emission time sequence during flywheel operation includes: Acquire acoustic emission signals; The acoustic emission signal is bandpass filtered; The filtered acoustic emission signal is segmented using a fixed sliding time window; Data blocks are generated based on the data points contained within each time window; Generate an acoustic emission time series based on all data blocks.

[0007] According to a specific implementation of this disclosure, calculating the fractal dimension of the acoustic emission signal based on the acoustic emission time series includes: The acoustic emission time series is normalized, including time coordinate normalization and signal amplitude normalization; Normalized time coordinates for: ; in, For indexes of time points, ; The total number of acoustic emission signal data points within a time window; Normalized signal amplitude for: ; in, The original amplitude of the data points that need to be normalized; This is a time series of acoustic emissions. This is the minimum value in the entire acoustic emission signal sequence; This is the maximum value in the entire acoustic emission signal sequence; The normalized acoustic emission time series is represented as a signal curve within a unit square [0,1]×[0,1]: ; With side length A square grid covers the entire unit square [0,1]×[0,1], and the total number of non-repeating grids crossed by the signal curve is counted. ;in, , This refers to the level number of the box size; Calculate the fractal dimension using the following formula. : ; in, A positive constant; N The total number of unique grid cells; Determine the fractal dimension Check if it is between [1,2]; if it is not in that range, recalculate.

[0008] According to a specific implementation of this disclosure, the calculation of the Lyapunov exponent based on the acoustic emission time series includes: Perform phase space reconstruction of the acoustic emission time series; In the reconstructed phase space, the initial neighboring points of each point are obtained to obtain the initial neighboring point pairs; The average divergence curve of the initial neighbor pairs as time progresses is calculated based on the following formula: ; in, The average divergence curve; For valid neighbor pairs Quantity; For all valid neighbor pairs Find the average; For the initial neighbor pairs of future Euclidean distance after one time step; The initial nearest neighbor pairs are calculated based on the following formula. Euclidean distance after one time step : ; in, The evolutionary position of the reference point; The evolutionary position of neighboring points; For the number of evolutionary steps, , ,and distance Not exceeding the diameter of the phase space attractor; Plot the mean divergence curve Follow A changing curve; Select the curve The data points within the linear interval that meet the preset criteria are fitted with least-squares linear data to obtain a straight line. :

[0009] in, To fit a straight line The slope; To fit the straight line in Intercept at 0; For the number of evolutionary steps

[0010] The Lyapunov index is calculated based on the following formula. : ; in, The slope of the fitted line; This represents the sampling time interval.

[0011] According to a specific implementation of this disclosure, calculating a health index based on the fractal dimension and the Lyapunov index includes: Normalize the fractal dimension and Lyapunov exponent: The health index is calculated based on the following formula. : ; in, , A positive constant; The weighting coefficients for the fractal dimension; The weighting coefficients of the Lyapunov index; This is the normalized offset of the fractal dimension; Normalized offset of the Lyapunov exponent

[0012] According to a specific implementation of an embodiment of this disclosure, predicting remaining lifespan based on the health index includes: Monitor the health index curve, and when no inflection point is detected, calculate the remaining life expectancy before the inflection point based on the following formula. :

[0013] in, For linear models The time it takes to reach the failure threshold; The current moment; When an inflection point is detected, the remaining lifetime after the inflection point is calculated based on the following formula. : ; in, For the rate of damage deterioration Current health index; This is the failure threshold.

[0014] Secondly, embodiments of this disclosure provide a flywheel life prediction system, the system comprising: The data acquisition module is configured to collect the acoustic emission time series during flywheel operation; The calculation module is configured to calculate the fractal dimension of the acoustic emission signal based on the acoustic emission time series; and to calculate the Lyapunov index based on the acoustic emission time series; and to calculate a health index based on the fractal dimension and the Lyapunov index. A lifespan prediction module is configured to predict remaining lifespan based on the health index.

[0015] Thirdly, embodiments of this disclosure also provide an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, cause the at least one processor to perform the flywheel life prediction method as described in the first aspect or any implementation thereof.

[0016] Fourthly, embodiments of this disclosure also provide a non-transitory computer-readable storage medium storing computer instructions that, when executed by at least one processor, cause the at least one processor to perform the flywheel life prediction method described in the first aspect or any implementation thereof.

[0017] Fifthly, embodiments of this disclosure also provide a computer program product, the computer program product including a computing program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the flywheel life prediction method in the first aspect or any implementation thereof.

[0018] The flywheel life prediction method disclosed in this embodiment analyzes the flywheel's emission signals to extract two key nonlinear features: the fractal dimension and the maximum Lyapunov exponent, and then uses these to predict life. This invention offers advantages such as strong early warning capabilities, high prediction accuracy, and strong anti-interference capabilities. It is extremely sensitive to microscopic damage within materials (such as microcrack initiation), and can identify critical points of damage and instability before macroscopic performance degradation or abnormal vibration occurs, allowing sufficient time for maintenance decisions. Furthermore, this method is not only applicable to flywheel systems; its principles based on acoustic emission and nonlinear analysis can also be extended to predictive maintenance of other rotating machinery such as bearings and gearboxes, showing broad application prospects. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a method for predicting the lifespan of a flywheel, provided in an embodiment of this disclosure. Figure 2 A flowchart illustrating a method for predicting the lifespan of a flywheel, as provided in this embodiment of the disclosure; Figure 3 This is a schematic diagram of a flywheel life prediction system provided in an embodiment of the present disclosure; Figure 4 A schematic diagram of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0020] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0021] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0022] It should be noted that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. Furthermore, this device and / or this method can be implemented using other structures and / or functionalities besides one or more of the aspects set forth herein.

[0023] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0024] This invention provides a method for predicting flywheel lifespan. By analyzing the fractal dimension and nonlinear dynamic characteristics of acoustic emission signals, the remaining lifespan of the flywheel can be predicted in real time. The acoustic emission signal is treated as a time series of a nonlinear dynamic system. By calculating its fractal dimension and Lyapunov exponent, the inherent complexity of damage evolution is quantified, thereby enabling early prediction of failure points.

[0025] Figure 1 This is a schematic diagram of the flywheel life prediction method provided in an embodiment of this disclosure.

[0026] Figure 2 To and Figure 1 The corresponding flowchart for the flywheel life prediction method.

[0027] like Figure 1 As shown, in step S110, the acoustic emission time sequence during the operation of the flywheel is collected.

[0028] When the flywheel is running, internal micro-damage (such as micro-cracks and friction) will release high-frequency stress waves.

[0029] High-frequency acoustic emission sensors (>100kHz) are installed at key locations in the flywheel bearing housing or hub to continuously acquire acoustic emission signals at a high sampling rate (e.g., 1MHz). Subsequent preprocessing includes bandpass filtering to remove mechanical and electronic noise, and segmentation of the signal using a sliding time window, assigning a representative time window to each segment. Dataset.

[0030] In this embodiment of the invention, the acquisition of the acoustic emission time series during flywheel operation includes: acquiring acoustic emission signals; performing bandpass filtering on the acoustic emission signals; segmenting the filtered acoustic emission signals through a fixed sliding time window; generating data blocks based on the data points contained in each time window; and generating an acoustic emission time series based on all data blocks.

[0031] More specifically, first, select a high-sensitivity acoustic emission sensor with a resonant frequency of 100-400kHz or a wide frequency range (such as 100kHz-1MHz).

[0032] Mount the acoustic emission sensor in the area closest to the potential damage source, such as the bearing housing of the main bearing, the flywheel hub, critical support structures, or other structures deemed necessary. Use a high-frequency coupling agent (such as silicone grease, a special adhesive, etc.) between the sensor and the flywheel surface to ensure efficient transmission of acoustic energy to the sensor. Securely fix the sensor using a magnetic mount, clamp, or adhesive to ensure stable coupling throughout operation; if a magnetic mount is used, it must be grounded to avoid loop current interference.

[0033] Secondly, the sensor signal is connected to a preamplifier and the gain is set to 20-40dB to improve the signal-to-noise ratio and facilitate transmission.

[0034] Choose a high-resolution (at least 16-bit), high-sampling-rate acquisition card. The sampling rate should be at least 2.5 times the sensor's highest effective frequency. For example, for a 300kHz sensor, the sampling rate should be set to at least 3-5 MS / s to avoid aliasing distortion. Enable the hardware anti-aliasing filter on the acquisition card, with its cutoff frequency set slightly below half the sampling rate.

[0035] The continuous stream acquisition mode is adopted, and the data is stored in a streaming manner in the form of continuous data blocks (such as data per second or per minute stored as a file).

[0036] Secondly, the acquired raw signals Digital bandpass filtering is performed to obtain a clear acoustic emission signal within a certain frequency band. .

[0037] Finally, the acoustic emission signal Segment and format.

[0038] Define a fixed sliding time window (the default sliding time window is 10 minutes, which can be adjusted according to the actual scenario) to divide the continuous acoustic emission data stream into continuous data segments. Each time window... (No. All data points contained within a 10-minute data segment constitute a data block, and all data blocks together form an acoustic emission time series, denoted as... .

[0039] More specifically, we now proceed to step S120.

[0040] In step S120, the fractal dimension of the acoustic emission signal is calculated based on the acoustic emission time series.

[0041] The evolution of material damage follows a typical fractal process, progressing from single microcracks to complex crack networks. Fractal dimension. It can effectively quantify the complexity of damage patterns. Early damage... Relatively stable, when damage begins to interconnect and expand rapidly, It will show a trend change.

[0042] In this embodiment of the invention, calculating the fractal dimension of the acoustic emission signal based on the acoustic emission time series includes: The acoustic emission time series is normalized, including time coordinate normalization and signal amplitude normalization; Normalized time coordinates for: ; in, For indexes of time points, ; The total number of acoustic emission signal data points within a time window; Normalized signal amplitude for: ; in, The original amplitude of the data points that need to be normalized; This is a time series of acoustic emissions. This is the minimum value in the entire acoustic emission signal sequence; This is the maximum value in the entire acoustic emission signal sequence; The normalized acoustic emission time series is represented as a signal curve within a unit square [0,1]×[0,1]: ; With side length A square grid covers the entire unit square [0,1]×[0,1], and the total number of non-repeating grids crossed by the signal curve is counted. ;in, , This refers to the level number of the box size; Calculate the fractal dimension using the following formula. : ; in, It is a positive constant and is related to the initial size or proportion of the fractal structure.

[0043] More specifically, the preprocessed acoustic emission signal waveform is quantized into fractal dimension. .

[0044] First, to eliminate the influence of signal amplitude and duration and make signals from different time windows comparable, the acoustic emission time series is... Normalize. Include discrete data points , .

[0045] Map the time axis (sampling point index) to the interval [0,1], and the normalized time coordinates are obtained. for:

[0046] Map the signal amplitude to the interval [0,1], and the normalized amplitude. for:

[0047] in, The original amplitude of the data points that need to be normalized; This is a time series of acoustic emissions. The acoustic emission signal is represented as a curve within a unit square [0,1]×[0,1]:

[0048] Secondly, the box counting method is applied, using a grid of boxes of different sizes to cover the signal curve, and the number of boxes through which the curve passes is counted.

[0049] Choose a sequence of square box sizes that gradually decrease. (A sequence decreasing from base 2) to ensure a sufficient number of points on a log-log coordinate system for linear fitting:

[0050] in, In the first The specific box size used at each level; This refers to the level number of the box size. , Satisfy the minimum The corresponding box can still cover multiple data points ( This helps to avoid excessive discretization errors.

[0051] Count the total number of unique boxes that are crossed by the signal curve. : With side length A square grid covers the entire unit square [0,1]×[0,1]. The number of rows and columns of the grid are both [0,1] and [0,1]. (Round up).

[0052] Calculate two adjacent data points and The line segment formed is counted, and the total number of non-repeating boxes crossed by the signal curve is denoted as . (The same box is counted only once).

[0053] Next, the fractal dimension is determined based on the box count results. .

[0054] For fractal curves, the total number of covered boxes With box size Satisfies the power law relationship:

[0055] Taking the logarithm of both sides, we get:

[0056] is a positive constant, and is the intercept in the linear fitting equation.

[0057] Right now, right In the graph, the data points should be approximately distributed along a straight line, and the absolute value of the slope of this line is the fractal dimension. By performing a linear fit on the data points using the least squares method, the intercept of the fitted line is... .

[0058] For each box size ,calculate:

[0059]

[0060] in, The x-axis variable is for linear fitting; The ordinate variable is for linear fitting; Box dimensions; The total number of boxes covered; A series of data points were obtained .

[0061] Use the least squares method on the data point set Perform linear fitting to obtain the best-fit line. ,in, for The calculated value; The slope; for The calculated value; For intercept (when =0 (value).

[0062] fractal dimension That is, the slope absolute value:

[0063] Finally, a validity test is conducted.

[0064] Determine the fractal dimension If the range is between 1 (a perfectly smooth straight line) and 2 (a curve that fills the entire plane), check the data and recalculate if it is not in that range.

[0065] Next, proceed to step S130.

[0066] In step S130, the Lyapunov exponent is calculated based on the acoustic emission time series.

[0067] In this embodiment of the invention, the calculation of the Lyapunov exponent based on the acoustic emission time series includes: Perform phase space reconstruction of the acoustic emission time series; In the reconstructed phase space, the initial neighboring points of each point are obtained to obtain the initial neighboring point pairs; The average divergence curve of the initial neighbor pairs as time progresses is calculated based on the following formula: ; in, The average divergence curve; For valid neighbor pairs Quantity (valid) satisfy , (Total points) For all valid neighbor pairs Find the average; For the initial neighbor pairs of future Euclidean distance after one time step; The initial nearest neighbor pairs are calculated based on the following formula. Euclidean distance after one time step : ; in, The evolutionary position of the reference point; The evolutionary position of neighboring points; For the number of evolutionary steps, , ,and distance Not exceeding the diameter of the phase space attractor; Plot the mean divergence curve Follow A changing curve; Select the curve The data points within the linear interval that meet the preset criteria are fitted with least-squares linear data to obtain a straight line. : ; in, To fit a straight line The slope; To fit the straight line in Intercept at 0; This represents the number of evolutionary steps.

[0068] The Lyapunov index is calculated based on the following formula. : ; in, The slope of the fitted line; This represents the sampling time interval.

[0069] More specifically, first, the maximum Lyapunov exponent of the acoustic emission signal is calculated. Maximum Lyapunov index A positive value and a larger value indicate a more chaotic system, where small initial differences will amplify exponentially over time.

[0070] Phase space reconstruction includes the following steps: Acquiring acoustic emission time series ,Include One data block.

[0071] Calculate the autocorrelation function of a time series. And select the autocorrelation function to decrease to the initial value for the first time. of (in, It is a constant. The delay corresponding to approximately 37% (Number of sampling points, i.e., optimal) The value will become the delay time. ).

[0072]

[0073] in, The length of the time series; For the number of evolutionary steps; This represents the sample mean of the time series. ; For time indexing; For time series in The value at any given moment.

[0074] Determining the embedding dimension based on the spurious nearest neighbor method Gradually increase the embedding dimension ,from Begin. In In 3D space, for each point Find its nearest neighbor Then look at these two points. Are they still nearest neighbors in the dimensional space? If so, with the increase of the embedding dimension... As the proportion of false nearest neighbors increases, it decreases to near 0 (when it is below 5%, it is considered close to 0), and then it is considered that at this time... It is a suitable embedding dimension.

[0075] Using a defined delay and embedding dimension One-dimensional time series Reconstructed dimensional phase space vector :

[0076] in, ; To reconstruct the total number of points in the phase space; For the original time series data points; each vector This represents the system in A state point in the phase space.

[0077] Secondly, in the reconstructed phase space, a reference trajectory that evolves over time and originates from its neighboring points is found for each point.

[0078] 1. Traverse the points in phase space ( ).

[0079] 2. Find the corresponding element in the entire phase space. The other point closest to Euclidean distance ( ),and ,in The minimum time interval (the average period of the flywheel rotation) is used to avoid selecting time-dependent spurious nearest neighbors.

[0080] 3. Record the initial distance between these two nearest neighbor points. .

[0081] Next, quantify the average rate at which this pair of initially neighboring points diverges over time.

[0082] 1. Track each pair of neighboring points In the future The evolution after a time step. That is, the calculation point. and points Euclidean distance between : ; in, The evolutionary position of the reference point; The evolutionary position of neighboring points; For the number of evolutionary steps, , ,and distance Not exceeding the diameter of the phase space attractor; 2. Calculate the average divergence: for each evolution step... Calculate all valid point pairs of Average value: ; in, The average divergence curve; For valid neighbor pairs Quantity (valid) satisfy , Total points ; For all valid neighbor pairs Find the average; Initial neighbor pairs for future Euclidean distance (mean divergence curve) after a time step.

[0083] Finally, based on the average divergence curve Get the maximum Lyapunov exponent .

[0084] 1. Drawing Follow Identify the most suitable linear interval from the changing curve. .

[0085] The most suitable linear interval can be identified manually or by algorithm, and has the following characteristics: a clear linear growth trend, a sufficiently long interval range (at least 10-15 consecutive values), a high goodness of fit (when performing a linear fit on the interval, the coefficient of determination should be very close to 1, for example, a coefficient of determination greater than 0.95), and avoidance of obvious nonlinear regions.

[0086] 2. For linear intervals Data points within By performing least-squares linear fitting, a straight line is obtained. ,in, For the number of evolutionary steps; To fit a straight line The slope represents Follow The average rate of change; To fit the straight line in The intercept at that point.

[0087] 3. Maximum Lyapunov index for: ; in, The slope of the fitted line; This represents the sampling time interval.

[0088] Will If considered as one unit of time, then , This represents the average rate of information loss or dispersion.

[0089] Next, proceed to step S140.

[0090] In step S140, a health index is calculated based on the fractal dimension and the Lyapunov index.

[0091] In this embodiment of the invention, calculating the health index based on the fractal dimension and the Lyapunov index includes: Normalize the fractal dimension and Lyapunov exponent: The health index is calculated based on the following formula. : ; in, , A positive constant; This is the normalized offset of the fractal dimension; This is the normalized offset of the Lyapunov exponent; The weighting coefficients for the fractal dimension; The weighting coefficients of the Lyapunov index.

[0092] More specifically, the fractal dimension (Geometric complexity) and Lyapunov exponent (Dynamic instability) is fused to obtain a comprehensive health index, and the critical point of the flywheel system transitioning from stable degradation to accelerated failure is identified based on the evolution trajectory of this index.

[0093] First, to eliminate the baseline value deviation caused by sensor differences and individual flywheel differences, and to make the health index comparable across different flywheel units, the following steps are taken: and Normalize: Use data from the initial healthy operation phase of the flywheel (the first 100 hours, or when flywheel health is confirmed). and The data calculates the baseline mean and standard deviation.

[0094]

[0095]

[0096]

[0097]

[0098] in, The baseline mean of the fractal dimension; The baseline standard deviation of the fractal dimension; The baseline mean of the Lyapunov index; is the baseline standard deviation of the Lyapunov index; The time frame for the healthy baseline period; It is a mean function; This is the standard deviation function.

[0099] The normalized offset is calculated based on the following formula:

[0100]

[0101] in, This is the standardized offset for the current fractal dimension; Let f(x) be the fractal dimension at the current moment. The healthy baseline mean of the fractal dimension; The baseline standard deviation of the fractal dimension; This is the standardized offset of the current Lyapunov exponent; The Lyapunov exponent at the current moment; The healthy baseline average of the Lyapunov index; is the baseline standard deviation of the Lyapunov index.

[0102] and It intuitively represents how many times the standard deviation of the current feature value deviates from the healthy baseline.

[0103] The health index is calculated based on the following formula. : A reciprocal fusion model is used to ensure that deterioration of any feature will lead to a decrease in the health index:

[0104] in, , A positive constant is used to ensure that the value within the parentheses is positive. For ease of calculation, this invention preferably uses 1 to ensure that the value within the parentheses is positive and to avoid issues arising in the early stages of health. The denominator is too small due to the value being close to 0 or negative. Abnormally large; The weighting coefficients for the fractal dimension; The weighting coefficients for the Lyapunov index can be determined through regression analysis of historical failure data or expert experience. , The initial value, when there is no prior data, will be , The initial value is set to 1, and the health index is normalized.

[0105] In the initial healthy state, , , It can be normalized to 1.0.

[0106] As the injury progresses, and As the denominator increases, the denominator increases. Monotonous decline.

[0107] when The value dropped to the preset failure threshold. When the flywheel fails, it is considered to have failed. The failure threshold is obtained based on historical failure data, calculated by subtracting the standard deviation of the health index at the time of historical failure. If no historical data is available, it can be set to 0.2.

[0108] Then, accurately locate the inflection point (critical point) where the health index curve changes from a relatively gentle decline to a steep decline. ).

[0109] The critical point corresponds to the transformation of the damage mechanism from independent and stable accumulation to interconnected and mutually reinforcing unstable expansion.

[0110] Specific recognition algorithm: 1. Calculate the first and second derivatives: Calculate the health index using the central difference method. First derivative with respect to time (Degradation rate) and second derivative (Degradation acceleration).

[0111]

[0112]

[0113] in, For the next time point Health index; For the previous time point Health index; Δt represents the current time; Δt represents the time step (e.g., 10 minutes).

[0114] 2. Inflection point detection based on second derivative: The mathematical definition of an inflection point is the point on a curve where the concavity or convexity changes, that is, the point where the second derivative crosses zero.

[0115] Traverse the time series and find points that satisfy the following conditions. :

[0116] in, , .

[0117] The first point that satisfies the above conditions is marked as a candidate critical point. .

[0118] 3. Based on the first derivative Confirmation of the minimum point: The critical point also corresponds to the first derivative. The minimum point of (degradation rate) (i.e. the moment when the rate of decline is the fastest).

[0119] At the candidate critical point In a nearby window ( ), searching The local minimum point. The time corresponding to this minimum point is determined as the more precise critical point. .

[0120] 4. Delay strategy for critical point confirmation: Derivative-based methods are sensitive to noise and may generate false alarms. Therefore, when the algorithm first detects a candidate critical point... Instead of immediately confirming the diagnosis, they continued to observe the subsequent events. A time window (e.g.) Within 50 minutes , behavior The size can be set according to different accuracy requirements, and there is no limit here.

[0121] If during this observation period, If the decline continues rapidly and no new inflection point appears, then it is officially confirmed. Calculate the standard deviation of the first derivative of the health index during healthy flywheel operation. When the health index... When the instantaneous rate of decline exceeds twice the standard deviation, it is considered to be in a rapid decline phase.

[0122] Next, proceed to step S150.

[0123] In step S150, the remaining lifespan is predicted based on the health index.

[0124] In this embodiment of the invention, predicting remaining lifespan based on the health index includes: Monitor the health index curve, and when no inflection point is detected, calculate the remaining life expectancy before the inflection point based on the following formula. : ; in, For linear models The time it takes to reach the failure threshold; The current moment; When an inflection point is detected, the remaining lifetime after the inflection point is calculated based on the following formula. : ; in, For the rate of damage deterioration Current health index; This is the failure threshold.

[0125] More specifically, based on health index The evolution trajectory of the flywheel is dynamically predicted from the current moment. The remaining time until expiration.

[0126] First, before reaching the critical point ( ( ) remaining life prediction.

[0127] Before the critical point ( (i.e., continuous monitoring of health indices) The curve, as long as no clear inflection point is detected (i.e. If it is not recognized, it is considered to be before the critical point.

[0128] Health decline is linear in the short term. Using recent data... Data (e.g., from) up to the current moment Data, A linear regression can be performed using 1 / 5 of the total running time to fit a straight line. :

[0129] in, It is a straight line The slope of the curve is usually negative, indicating the rate of degradation. It is a straight line The intercept represents the initial fitted value of the health index of the linear model at the start time. This refers to the current moment.

[0130] Solving the linear model Reaching the failure threshold Time : At the point of failure, the health index equals the failure threshold, i.e. .

[0131]

[0132] The remaining lifetime before the critical point The prediction is:

[0133] in, For the current moment, .

[0134] Secondly, critical point identification and confirmation are performed. ).

[0135] Real-time computing The first and second derivatives. When the second derivative changes from negative to positive (the health index curve changes from convex to concave), this moment is initially marked as a candidate critical point. .

[0136] observe After a period of time (e.g., after another 5-10 data points) Data. If If it enters a sustained, accelerating downward channel, then it is confirmed. This is to avoid misjudgments caused by fluctuations in data noise.

[0137] After confirmation, the critical point time will be set. and the health index value at that point Record this as the starting point for the accelerated failure model.

[0138] Finally, after reaching the critical point ( ( ) remaining life prediction.

[0139] After the critical point, the evolution of damage is self-accelerating, and the decline in health index follows an exponential law.

[0140] The exponential decay model is as follows:

[0141] in, A health index at the critical point; The rate of damage deterioration; It is a constant.

[0142] Using from the critical point up to the current moment All health indicators The data is used to fit an exponential model. The model is linearized by taking the natural logarithm of both sides.

[0143] make , , Then the above equation becomes:

[0144] Use linear least squares to analyze the data points. The absolute value of the slope obtained by fitting the data is the rate of deterioration. The estimated value. Parameters The value of is dynamically updated as new data accumulates after the critical point, making the estimation results increasingly accurate.

[0145] Let the current time be The current health index is Calculate the health index from the current value. Exponential decay to failure threshold Time required :

[0146] Right now,

[0147] in, This is the time required for exponential decay to the failure threshold.

[0148] The remaining lifespan prediction at the current moment is:

[0149] The flywheel life prediction method proposed in this invention, through continuous monitoring, feature extraction, condition assessment, and model prediction, transforms raw, microscopic acoustic emission activity into macroscopic, quantitative engineering decision-making information (remaining life). This provides a critical time window for predictive maintenance of flywheel systems, thereby avoiding sudden failures, scheduling planned maintenance, and significantly improving the reliability and availability of flywheel systems.

[0150] Figure 3 The present invention illustrates a flywheel life prediction system 300, which includes a data acquisition module 310, a calculation module 320, and a life prediction module 330.

[0151] Data acquisition module 310 is used to acquire the acoustic emission time series during flywheel operation; The calculation module 320 is used to calculate the fractal dimension of the acoustic emission signal based on the acoustic emission time series; and to calculate the Lyapunov index based on the acoustic emission time series; and to calculate a health index based on the fractal dimension and the Lyapunov index. The lifespan prediction module 330 is used to predict the remaining lifespan based on the health index.

[0152] See Figure 4 This disclosure also provides an electronic device 40, which includes: At least one processor; and, The memory is communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the flywheel life prediction method in the foregoing method embodiments.

[0153] This disclosure also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the flywheel life prediction method described in the foregoing method embodiments.

[0154] This disclosure also provides a computer program product, which includes a computing program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions that, when executed by a computer, cause the computer to perform the flywheel life prediction method described in the foregoing method embodiments.

[0155] The following is for reference. Figure 4 The diagram illustrates a structural schematic of an electronic device 40 suitable for implementing embodiments of the present disclosure. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0156] like Figure 4 As shown, electronic device 40 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 402 or a program loaded from storage device 408 into random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of electronic device 40. The processing unit 401, ROM 402, and RAM 403 are interconnected via bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0157] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic device 40 to communicate wirelessly or wiredly with other devices to exchange data. Although electronic device 40 with various devices is shown in the figure, it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0158] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a storage device 408, or installed from a ROM 402. When the computer program is executed by the processing device 401, it performs the functions defined in the methods of embodiments of this disclosure.

[0159] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0160] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0161] The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire at least two Internet Protocol (IP) addresses; send a node evaluation request including the at least two IP addresses to a node evaluation device, wherein the node evaluation device selects an IP address from the at least two IP addresses and returns it; and receive the IP address returned by the node evaluation device; wherein the acquired IP address indicates an edge node in a content delivery network.

[0162] Alternatively, the aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: receive a node evaluation request including at least two Internet Protocol (IP) addresses; select an IP address from the at least two IP addresses; and return the selected IP address; wherein the received IP address indicates an edge node in the content delivery network.

[0163] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0164] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0165] The units described in the embodiments of this disclosure can be implemented in software or in hardware. The name of a unit does not necessarily limit the unit itself; for example, the first acquisition unit can also be described as "a unit that acquires at least two Internet Protocol addresses".

[0166] It should be understood that the various parts of this disclosure can be implemented in hardware, software, firmware, or a combination thereof.

[0167] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for predicting the lifespan of a flywheel, characterized in that, The method includes the following steps: Acoustic emission time series of the flywheel during operation; The fractal dimension of the acoustic emission signal is calculated based on the acoustic emission time series. Calculate the Lyapunov exponent based on the acoustic emission time series; The health index is calculated based on the fractal dimension and the Lyapunov index. Predict remaining life expectancy based on the aforementioned health index; The health index is calculated based on the fractal dimension and the Lyapunov index, including: Normalize the fractal dimension and Lyapunov exponent: The health index is calculated based on the following formula. : ； in, , A positive constant; The weighting coefficients for the fractal dimension; The weighting coefficients of the Lyapunov index; This is the normalized offset of the fractal dimension; This is the normalized offset of the Lyapunov exponent; Predicting remaining lifespan based on the aforementioned health index includes: Monitor the health index curve, and when no inflection point is detected, calculate the remaining life expectancy before the inflection point based on the following formula. : ； in, For linear models The time it takes to reach the failure threshold; The current moment; When an inflection point is detected, the remaining lifetime after the inflection point is calculated based on the following formula. : ； in, The rate of damage deterioration; Current health index; This is the failure threshold.

2. The method for predicting flywheel lifespan according to claim 1, characterized in that, The time series of acoustic emissions collected during the operation of the flywheel includes: Acquire acoustic emission signals; The acoustic emission signal is bandpass filtered; The filtered acoustic emission signal is segmented using a fixed sliding time window; Data blocks are generated based on the data points contained within each time window; Generate an acoustic emission time series based on all data blocks.

3. The method for predicting flywheel lifespan according to claim 1, characterized in that, The calculation of the fractal dimension of the acoustic emission signal based on the acoustic emission time series includes: The acoustic emission time series is normalized, including time coordinate normalization and signal amplitude normalization; Normalized time coordinates for: ； in, For indexes of time points, ; The total number of acoustic emission signal data points within a time window; Normalized signal amplitude for: ； in, The original amplitude of the data points that need to be normalized; This is a time series of acoustic emissions. This is the minimum value in the entire acoustic emission signal sequence; This is the maximum value in the entire acoustic emission signal sequence; The normalized acoustic emission time series is represented as a signal curve within a unit square [0,1]×[0,1]: ； With side length A square grid covers the entire unit square [0,1]×[0,1], and the total number of non-repeating grids crossed by the signal curve is counted. ;in, , This refers to the level number of the box size; Calculate the fractal dimension using the following formula. : ； in, A positive constant; N The total number of unique grid cells; Determine the fractal dimension Check if it is between [1,2]; if it is not in that range, recalculate.

4. The method for predicting flywheel life according to claim 1, characterized in that, The calculation of the Lyapunov exponent based on the acoustic emission time series includes: Perform phase space reconstruction of the acoustic emission time series; In the reconstructed phase space, the initial neighboring points of each point are obtained to obtain the initial neighboring point pairs; The average divergence curve of the initial neighbor pairs as time progresses is calculated based on the following formula: ； in, The average divergence curve; For valid neighbor pairs Quantity; For all valid neighbor pairs Find the average; For the initial neighbor pairs of future Euclidean distance after one time step; The initial nearest neighbor pairs are calculated based on the following formula. Euclidean distance after one time step : ； in, The evolutionary position of the reference point; The evolutionary position of neighboring points; For the number of evolutionary steps, , ,and distance Not exceeding the diameter of the phase space attractor; Plot the mean divergence curve Follow A changing curve; Select the curve The data points within the linear interval that meet the preset criteria are fitted with least-squares linear data to obtain a straight line. : ； in, To fit a straight line The slope; To fit the straight line in Intercept at 0; For the number of evolutionary steps; The Lyapunov index is calculated based on the following formula. : ； in, The slope of the fitted line; This represents the sampling time interval.

5. A flywheel life prediction system, characterized in that, The system includes: The data acquisition module is configured to collect the acoustic emission time series during flywheel operation; The calculation module is configured to calculate the fractal dimension of the acoustic emission signal based on the acoustic emission time series; and to calculate the Lyapunov exponent based on the acoustic emission time series; calculate a health index based on the fractal dimension and the Lyapunov exponent; the calculation of the health index based on the fractal dimension and the Lyapunov exponent includes: Normalize the fractal dimension and Lyapunov exponent: The health index is calculated based on the following formula. : ； in, , A positive constant; The weighting coefficients for the fractal dimension; The weighting coefficients of the Lyapunov index; This is the normalized offset of the fractal dimension; This is the normalized offset of the Lyapunov exponent; A lifespan prediction module is configured to predict remaining lifespan based on the health index. Predicting remaining lifespan based on the aforementioned health index includes: Monitor the health index curve, and when no inflection point is detected, calculate the remaining life expectancy before the inflection point based on the following formula. : ； in, For linear models The time it takes to reach the failure threshold; The current moment; When an inflection point is detected, the remaining lifetime after the inflection point is calculated based on the following formula. : ； in, The rate of damage deterioration; Current health index; This is the failure threshold.

6. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, cause the at least one processor to perform the flywheel life prediction method as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions that, when executed by at least one processor, cause the at least one processor to perform the flywheel life prediction method as described in any one of claims 1 to 4.

8. A computer program product, characterized in that, The computer program product includes a computing program stored on a non-transitory computer-readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform the flywheel life prediction method as described in any one of claims 1 to 4.

Images