Fan running state detection method and system, intelligent electric appliance and storage medium

Abstract

The invention provides a fan running state detection method and system, an intelligent electric appliance and a storage medium. The detection method comprises the following steps: acquiring a noise time domain signal and a vibration time domain signal of the fan when the fan is in an operation state of a maximum gear; extracting an acoustic characteristic value representing an abnormal sound prominence degree from the noise time domain signal, and generating a noise state score through a first preset dynamic scoring rule based on the acoustic characteristic value; extracting a vibration characteristic value representing the single-frequency vibration intensity from the vibration time domain signal, and generating a vibration state score through a second preset dynamic scoring rule based on the vibration characteristic value; according to the service life of the fan, calculating a service life correction factor through a segmented incremental function; and carrying out weighted fusion on the noise state score, the vibration state score and the age limit correction factor to generate a comprehensive state score. The comprehensive score is generated by fusing three elements of noise, vibration and age limit, so that the triple technical effects of reducing the maintenance cost, prolonging the service life of the fan and enabling the environmental noise to reach the standard are achieved.

Description

Technical Field

[0001] This disclosure relates to the field of smart appliances, and more particularly to a method, system, smart appliance, and storage medium for detecting the operating status of a fan. Background Technology

[0002] As people's living standards improve and technologies such as the internet, big data, artificial intelligence, and voice interaction become more widespread, traditional lifestyles are gradually changing, and the use of home appliances is increasingly moving towards intelligentization. While bringing more convenience to users, the functions of various home appliances are also becoming more diversified.

[0003] Currently, the condition assessment of outdoor fans in commercial range hoods mainly relies on users' subjective perceptions or fixed-cycle maintenance plans, lacking objective quantitative evidence. While existing technologies can detect single parameters (such as total sound pressure level or total vibration), they haven't established a correlation model between age and performance degradation, leading to the use of the same evaluation standards for both new and old equipment. This makes it difficult for users to accurately assess the fan's true condition, easily resulting in problems such as untimely maintenance (shortened lifespan), premature replacement (waste of resources), or continuous operation exceeding standards (nuisance and hazard to residents). Summary of the Invention

[0004] The technical problem to be solved by this disclosure is to overcome the deficiency in the existing technology of not having established a dynamic quantitative evaluation system that integrates noise abnormality characteristics, vibration frequency energy and equipment aging factors in order to achieve accurate classification decision on the condition of the fan, and to provide a method, system, intelligent electrical appliance and storage medium for detecting the operating condition of the fan.

[0005] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0006] This disclosure provides a method for detecting the operating status of a wind turbine, the method comprising:

[0007] With the fan operating at its maximum speed, the time-domain noise signal and the time-domain vibration signal of the fan are collected.

[0008] Acoustic feature values ​​representing the prominence of abnormal sounds are extracted from the noise time-domain signal, and a noise state score is generated based on the acoustic feature values ​​through a first preset dynamic scoring rule;

[0009] Vibration feature values ​​characterizing the intensity of single-frequency vibration are extracted from the vibration time-domain signal, and a vibration state score is generated based on the vibration feature values ​​through a second preset dynamic scoring rule;

[0010] Based on the service life of the wind turbine, the service life correction factor is calculated using a piecewise increasing function.

[0011] The noise status score, the vibration status score, and the age correction factor are weighted and fused to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

[0012] Optionally, extracting acoustic feature values ​​characterizing the prominence of abnormal sounds from the noise time-domain signal includes:

[0013] The noise time-domain signal is converted into a frequency domain to obtain the sound pressure spectrum;

[0014] Extract the maximum sound pressure level peak value of the sound pressure spectrum;

[0015] Calculate the effective sound pressure level for the entire time period based on the noise time-domain signal;

[0016] The difference between the effective sound pressure level and the peak value of the maximum sound pressure level is used as the acoustic characteristic value.

[0017] Optionally, extracting the maximum sound pressure level peak value of the sound pressure spectrum includes: identifying abnormal frequency bands in the sound pressure spectrum where the gradient change rate is greater than or equal to a change rate threshold, and extracting the maximum sound pressure level peak value from the abnormal frequency bands;

[0018] And / or,

[0019] The step of calculating the effective sound pressure level over the entire time period based on the noise time-domain signal includes: calculating the root mean square value of the sound pressure for all sampling points of the noise time-domain signal, and converting the root mean square value of the sound pressure into a decibel value as the effective sound pressure level.

[0020] Optionally, the step of extracting vibration feature values ​​characterizing the intensity of single-frequency vibration from the vibration time-domain signal includes:

[0021] The vibration time-domain signal is converted into a frequency domain to obtain the vibration spectrum;

[0022] Extract the maximum vibration peak value of the vibration spectrum;

[0023] Calculate the effective vibration value for the entire time period based on the vibration time-domain signal;

[0024] The ratio of the maximum vibration peak value to the effective vibration value is used as the vibration characteristic value.

[0025] Optionally, extracting the maximum vibration peak value of the vibration spectrum includes: extracting the point with the largest amplitude value as the characteristic frequency peak value within the preset frequency range of the fan;

[0026] And / or,

[0027] The calculation of the effective vibration value for the entire time period based on the vibration time-domain signal includes: calculating the root mean square value of acceleration for all sampling points of the vibration time-domain signal, and using the root mean square value of acceleration as the effective vibration value.

[0028] Optionally, the weighted fusion satisfies at least one of the following:

[0029] The first contribution weight of the noise state score is greater than the second contribution weight of the vibration state score;

[0030] The age correction factor increases monotonically with the service life.

[0031] The output value of the weighted fusion is the product of the age correction factor and the weighted score, wherein the weighted score includes: the product of the noise state score and the first contribution weight, and the product of the vibration state score and the second contribution weight.

[0032] This disclosure provides a system for detecting the operating status of a wind turbine, the system comprising:

[0033] The acquisition module is used to acquire the noise time-domain signal and vibration time-domain signal of the fan when the fan is running at its maximum speed.

[0034] The first scoring module is used to extract acoustic feature values ​​representing the prominence of abnormal sounds from the noise time-domain signal, and generate a noise state score based on the acoustic feature values ​​through a first preset dynamic scoring rule.

[0035] The second scoring module is used to extract vibration feature values ​​characterizing the intensity of single-frequency vibration from the vibration time-domain signal, and generate a vibration state score based on the vibration feature values ​​through a second preset dynamic scoring rule.

[0036] The service life correction module is used to calculate the service life correction factor based on the service life of the wind turbine using a piecewise increasing function.

[0037] The third scoring module is used to weight and fuse the noise status score, the vibration status score, and the age correction factor to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

[0038] Optionally, the first scoring module is specifically used for:

[0039] The noise time-domain signal is converted into a frequency domain to obtain the sound pressure spectrum;

[0040] Extract the maximum sound pressure level peak value of the sound pressure spectrum;

[0041] Calculate the effective sound pressure level for the entire time period based on the noise time-domain signal;

[0042] The difference between the effective sound pressure level and the peak value of the maximum sound pressure level is used as the acoustic characteristic value.

[0043] Optionally, the first scoring module is specifically used to: identify abnormal frequency bands in the sound pressure spectrum where the gradient change rate is greater than or equal to a change rate threshold, and extract the peak value of the maximum sound pressure level from the abnormal frequency bands;

[0044] And / or,

[0045] The first scoring module is specifically used to: calculate the root mean square value of sound pressure for all sampling points of the noise time-domain signal, and convert the root mean square value of sound pressure into a decibel value as the effective sound pressure level.

[0046] Optionally, the second scoring module is specifically used for:

[0047] The vibration time-domain signal is converted into a frequency domain to obtain the vibration spectrum;

[0048] Extract the maximum vibration peak value of the vibration spectrum;

[0049] Calculate the effective vibration value for the entire time period based on the vibration time-domain signal;

[0050] The ratio of the maximum vibration peak value to the effective vibration value is used as the vibration characteristic value.

[0051] Optionally, the second scoring module is specifically used to: extract the maximum amplitude point as the characteristic frequency peak value within the preset frequency range of the wind turbine;

[0052] And / or,

[0053] The second scoring module is specifically used to: calculate the root mean square value of acceleration for all sampling points of the vibration time-domain signal, and the root mean square value of acceleration is used as the effective vibration value.

[0054] Optionally, the weighted fusion satisfies at least one of the following:

[0055] The first contribution weight of the noise state score is greater than the second contribution weight of the vibration state score;

[0056] The age correction factor increases monotonically with the service life.

[0057] The output value of the weighted fusion is the product of the age correction factor and the weighted score, wherein the weighted score includes: the product of the noise state score and the first contribution weight, and the product of the vibration state score and the second contribution weight.

[0058] This disclosure provides a smart appliance, including a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the method for detecting the operating status of a fan as described above.

[0059] This disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for detecting the operating status of a wind turbine as described above.

[0060] This disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the method for detecting the operating status of a wind turbine as described above.

[0061] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0062] The positive and progressive effects of this disclosure are as follows: by integrating the three factors of noise, vibration and age to generate a comprehensive score, it achieves a triple technical effect of reducing maintenance costs (reducing over-maintenance), extending the life of the fan (avoiding malfunctions), and ensuring that environmental noise meets standards (timely replacement of equipment exceeding standards). Attached Figure Description

[0063] Figure 1 A flowchart illustrating a method for detecting the operating status of a wind turbine, provided as an exemplary embodiment of this disclosure;

[0064] Figure 2 A flowchart of step 102 provided for an exemplary embodiment of this disclosure;

[0065] Figure 3 A flowchart of step 103 provided for an exemplary embodiment of this disclosure;

[0066] Figure 4 A schematic diagram of a wind turbine operating status detection system provided as an exemplary embodiment of this disclosure;

[0067] Figure 5 This is a schematic diagram of the structure of a smart appliance provided as an exemplary embodiment of the present disclosure. Detailed Implementation

[0068] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0069] The prefixes such as "first" and "second" used in this disclosure are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this disclosure does not constitute a limitation on the described objects. The description of the described objects is given in the context of the embodiments, and the use of such prefixes should not constitute unnecessary restrictions. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0070] In this embodiment of the disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information comply with relevant laws and regulations and do not violate public order and good morals.

[0071] Example 1

[0072] Figure 1 A flowchart illustrating a method for detecting the operating status of a wind turbine, provided as an exemplary embodiment of this disclosure, is included. The detection method comprises:

[0073] Step 101: With the fan operating at its maximum speed, collect the time-domain noise signal and the time-domain vibration signal of the fan.

[0074] Step 101 is the basic data acquisition stage for wind turbine condition monitoring. Its core is to acquire raw physical signals under the most demanding operating conditions of the wind turbine. The reason for selecting the maximum operating speed is:

[0075] 1. Representative operating conditions: At the maximum speed, the fan load is the highest, and potential fault characteristics such as mechanical wear, impeller imbalance, and bearing aging will be significantly amplified;

[0076] 2. Signal saliency: Abnormal noise and vibration have more concentrated energy at high speeds, which facilitates subsequent feature extraction;

[0077] 3. Consistency of testing: Avoid data incomparability due to the randomness of gear positions and ensure consistent evaluation standards.

[0078] The key points of implementation are explained in detail below:

[0079] 1. Operation status control

[0080] Fan start-up requirements: Ensure the fan has entered a stable operating state (usually requires 3-5 minutes of preheating) and avoid collecting transient interference signals during startup;

[0081] Gear confirmation mechanism: Verify whether the rated maximum speed has been reached through the gear feedback signal of the fan control system or the speed sensor.

[0082] 2. Signal Acquisition Standards

[0083] (a) Noise time-domain signal acquisition

[0084] Sensing device: A Class 1 sound level meter conforming to IEC 61672 standard is used, with a frequency response range covering 20Hz-20kHz (the audible range of the human ear).

[0085] Installation location: The microphone should be pointed at the center axis of the fan outlet, 1 meter away (to avoid near-field turbulence interference), and 1.5 meters above the ground (simulating human ear height).

[0086] Example of sampling requirements:

[0087] Sampling frequency ≥ 16kHz (satisfies Shannon's sampling theorem, fully preserving high-frequency abnormal sound components);

[0088] A-weighted network activation (simulating the human ear's sensitivity to frequency).

[0089] Sampling duration ≥ 5 seconds (covering at least 50 impeller rotation cycles);

[0090] Example: A commercial range hood fan has a rated maximum speed of 1500 RPM and 12 impeller blades. Therefore, the impeller passing frequency is 1500 / 60×12=300Hz. 5 seconds can cover 1500 cycles, which can fully characterize periodic abnormal noise.

[0091] (b) Vibration time-domain signal acquisition

[0092] Sensing device: ICP type triaxial accelerometer with a range of ≥50g (covering the maximum vibration and impact of the fan).

[0093] Installation location: Rigidly installed at the connection between the fan base and the ground, using honey grease bonding to ensure a high frequency response of over 5kHz;

[0094] Example of sampling requirements:

[0095] Sampling frequency ≥ 5kHz (meets 10 times the sampling rate of the highest characteristic frequency of the wind turbine, such as the bearing failure frequency of about 2kHz).

[0096] Axial selection: Prioritize the vertical direction (Z-axis), because 70% of the fan's vibration energy is concentrated in the vertical direction.

[0097] Three verifications need to be performed after data collection:

[0098] 1. Time-domain waveform integrity: Visually inspect the waveform for any breaks, saturation, or zero drift;

[0099] 2. Frequency domain energy distribution: The FFT spectrum should be continuous and without gaps in the 0-5kHz range;

[0100] 3. Statistical characteristic verification: Calculate the peak factor (Crest Factor) = peak value / RMS value. The normal range is 3-5. If it is >10, it indicates that the impact event has interfered and resampling is required.

[0101] This step is compatible with different fan types:

[0102] Centrifugal fans: Noise collection at the air inlet needs to be added (the main source of eddy current noise);

[0103] Axial flow fan: Vibration measurement points are added at the non-drive end of the motor (sensitive point for bearing failure);

[0104] Variable frequency fan: Real-time speed needs to be recorded synchronously (by reading the inverter output frequency via Modbus).

[0105] The time-domain signal output in step 101 directly determines the reliability of subsequent analyses:

[0106] Noise signal → Frequency domain transformation in step 102 → Abnormal frequency band identification → Acoustic feature extraction;

[0107] Vibration signal → frequency harmonic analysis in step 103 → generation of vibration characteristic values.

[0108] Step 102: Extract acoustic feature values ​​representing the prominence of abnormal sounds from the noise time-domain signal, and generate a noise state score based on the acoustic feature values ​​through the first preset dynamic scoring rule;

[0109] Step 102 is the core of the wind turbine noise status assessment. Its innovation lies in transforming the subjectively perceived "abnormal noise" into quantifiable parameters, breaking through the limitations of traditional methods that rely solely on total sound pressure level. By extracting acoustic feature values ​​and applying dynamic scoring rules, the following is achieved:

[0110] 1. Precise quantification of abnormal sounds: Capture discrete abnormal sounds caused by impeller damage, bearing dry friction, etc. (usually masked by background noise);

[0111] 2. Dynamic mapping of status scores: Adaptively generate scores based on the severity of abnormal sounds to avoid misjudgments caused by fixed thresholds.

[0112] Optionally, see Figure 2 It can be seen that step 102, extracting acoustic feature values ​​representing the prominence of abnormal sounds from the noise time-domain signal, includes:

[0113] Step 1021: Perform frequency domain transformation on the noise time-domain signal to obtain the sound pressure spectrum;

[0114] Step 1022: Extract the peak value of the maximum sound pressure level in the sound pressure spectrum;

[0115] Step 1023: Calculate the effective sound pressure level for the entire time period based on the noise time-domain signal;

[0116] Step 1024: Use the difference between the effective sound pressure level and the peak value of the maximum sound pressure level as the acoustic characteristic value.

[0117] The implementation details of the sub-steps are as follows:

[0118] Step 1021: Frequency domain transformation to obtain the sound pressure spectrum;

[0119] Technical objective: To decompose the time-domain sound pressure signal into a frequency-domain energy distribution, highlighting anomalous frequency bands;

[0120] The conversion methods include:

[0121] Basic scheme: Fast Fourier Transform (FFT), calculation formula:

[0122] ;

[0123] Where x(n) is the time-domain sampled value and N is the number of sample points.

[0124] Enhancement solution: Add a Hanning window to reduce spectral leakage; window function:

[0125] ;

[0126] Example of parameter requirements:

[0127] Frequency resolution ≤ 5Hz (to ensure precise location of abnormal frequency points);

[0128] Dynamic range ≥80dB (covering the fan from low-frequency roar to high-frequency howling);

[0129] Example: For a time-domain signal with a sampling rate of 16kHz and 8192 points, the 0-8kHz spectrum is obtained after FFT, with a resolution of ≈2Hz, which can accurately locate the characteristic frequency of bearing damage (such as 4.2kHz).

[0130] Step 1022: Extract the peak value of the maximum sound pressure level L1;

[0131] Technical objective: To identify the most prominent anomalous energy concentration points in the spectrum;

[0132] Extraction strategy:

[0133] 1. Full-band scanning: Directly takes the global maximum value of the spectrum (suitable for strong abnormal noise scenarios);

[0134] 2. Abnormal frequency band location (preferred):

[0135] Step 1: Calculate the rate of change of the spectral gradient ;

[0136] Step 2: Mark The frequency band is an anomaly (sudden change region);

[0137] Step 3: Take the point with the highest sound pressure level in the abnormal frequency band as L1;

[0138] Engineering significance: The gradient thresholding method can eliminate gentle broadband noise (such as wind noise) and focus on steep narrowband noise;

[0139] Example: A wind turbine spectrum shows a bulge with a gradient change rate of 25dB / oct at 3200Hz, and the peak value at this frequency is L1=84dB.

[0140] Step 1023: Calculate the effective sound pressure level over the entire time period. ;

[0141] Technical objective: To characterize the average acoustic energy throughout the entire sampling period;

[0142] Calculation process:

[0143] 1. Root mean square value in the time domain:

[0144] ;

[0145] 2. Convert decibel values:

[0146] ;

[0147] Example: Measured ,but ;

[0148] Step 1024: Generate acoustic feature values ;

[0149] definition: (Unit: dB);

[0150] Physical meaning:

[0151] L s (e.g. L) s >11dB) → Abnormal peaks are masked by background noise → No abnormal noise;

[0152] L s Small value (e.g., L) s <8dB) → Abnormal peak value is significantly prominent → There is obvious abnormal sound;

[0153] Technical advantages: Compared to traditional total sound pressure levels, L s Sensitivity to unusual sounds increased by more than 3 times (experimental data);

[0154] The first preset dynamic scoring rule will Mapped to noise state score:

[0155] ;

[0156] Scoring logic:

[0157] Segmentation design basis: Nonlinear perception of the human ear's 8-11dB masking effect (Weber-Fechner law).

[0158] Project correspondence:

[0159] → Normal condition (no abnormal noise);

[0160] → Cleaning is recommended (slight noise).

[0161] → Need to be replaced (severe abnormal noise);

[0162] Application example: When L is detected s When =7dB, S noise =7+0.375×(8-7)=7.375 → Trigger the "Need to be cleaned" instruction message.

[0163] Typical application scenarios:

[0164] Commercial range hoods: Detect impeller noise caused by grease buildup (characteristic frequency band 2-4kHz);

[0165] Industrial centrifugal fans: Capture high-frequency whistling (>5kHz) caused by insufficient bearing lubrication;

[0166] Cooling tower fan: Identify the modulation sideband (fundamental frequency ± fault frequency) caused by blade cracks.

[0167] The noise condition score generated in this step will be used in conjunction with the vibration score and age factor for decision-making.

[0168] High noise score + low vibration score → Impeller fouling and clogging → Trigger cleaning instruction message;

[0169] High noise score + high vibration score → Severe bearing wear → Triggering replacement instruction message;

[0170] By integrating multiple parameters, the limitations of traditional single-parameter detection are overcome (such as high vibration alone may be misjudged as a dynamic balance problem rather than bearing damage).

[0171] Optionally, step 1022, extracting the maximum sound pressure level peak value of the sound pressure spectrum, includes: identifying abnormal frequency bands in the sound pressure spectrum where the gradient change rate is greater than or equal to the change rate threshold, and extracting the maximum sound pressure level peak value from the abnormal frequency bands;

[0172] And / or,

[0173] Step 1023: Calculate the effective sound pressure level for the entire time period based on the noise time-domain signal, including: calculating the root mean square value of sound pressure for all sampling points of the noise time-domain signal, and converting the root mean square value of sound pressure into a decibel value as the effective sound pressure level.

[0174] The core of step 1022 lies in accurately locating heterophonic frequency bands through spectral gradient analysis. Its scientific basis is the human ear's sensitivity to steep spectral changes (psychoacoustic principles). Compared to directly extracting global maxima, this method effectively eliminates broadband background noise interference and focuses on narrowband discrete components caused by mechanical faults.

[0175] Threshold setting basis: The gradient change rate reflects the energy concentration in the frequency band. The threshold is usually set at 15-25 dB / oct (octave), preferably 20 dB / oct. Its physical meaning is: when the sound pressure level increases by more than the threshold for every doubling of the frequency, it is judged as an abnormal bulge.

[0176] Example: If the sound pressure level in a certain frequency band increases by 22dB from 1000Hz to 2000Hz (1 octave), then the gradient change rate = 22dB / oct > 20dB / oct, and it is determined to be an abnormal frequency band.

[0177] Gradient calculation methods:

[0178] The local gradient is calculated using the three-point difference method:

[0179] (Unit: dB / Hz);

[0180] Then convert to dB / oct:

[0181] ;

[0182] Note: L i For frequency f i For sound pressure level, frequency resolution must be ≤5Hz to ensure accuracy;

[0183] Special scenario processing for abnormal frequency band peak extraction includes:

[0184] 1. Multi-peak competition: When there are multiple frequency bands that satisfy the gradient threshold, the frequency band with the highest absolute sound pressure level is selected;

[0185] Example: The peak value in the 1200-1500Hz band is 82dB, and the peak value in the 3000-3500Hz band is 78dB, so the 1200-1500Hz band is selected.

[0186] 2. Broadband noise interference: If the abnormal frequency band width is >1 / 3 octave, the narrowband characteristics (Q factor >10) need to be verified.

[0187] ;

[0188] Step 1023 clarifies the complete path for calculating sound pressure energy from the original time-domain signal, avoids frequency-domain inversion errors, and ensures that the results meet international standards.

[0189] Example of sampling point processing requirements:

[0190] It must include all valid sampling points, excluding transient signals during the start / stop phase;

[0191] Amplitude limiting protection: Set a hardware limiter (±5V) or digital clipping detection (resample when peak factor > 10).

[0192] Calculation principle:

[0193]

[0194] Where p(n) is the instantaneous sound pressure at the nth sampling point, and N is the total number of sampling points;

[0195] Example of metrological requirements for decibel conversion:

[0196] Reference sound pressure level definition: Strictly adopts 20 micropascals, which is the airborne sound reference value specified in the international standard ISO 80000-8.

[0197] Conversion formula:

[0198]

[0199] Metrological verification:

[0200] Acquire a 94dB@1kHz standard sound source signal and calculate L. eq The error must be <±0.5dB (IEC 61672 Class 1 accuracy requirement).

[0201] Connection with step 1021:

[0202] Frequency domain transformation (step 1021) and RMS value calculation (step 1023) must use the same time window to avoid energy non-conservation due to signal truncation;

[0203] Example: If the FFT uses a Hanning window, then The energy loss coefficient of the window function to be compensated is 1.5.

[0204] Logical closed loop with step 1024:

[0205] The physical significance lies in quantifying the prominence of anomalous energy relative to the background:

[0206] L1 comes from the abnormal frequency band located by the gradient threshold method → ​​ensuring the capture of mechanical fault characteristics;

[0207] L eq Time-domain computation ensures accurate representation of background noise.

[0208] Step 103: Extract vibration feature values ​​representing the intensity of single-frequency vibration from the vibration time-domain signal, and generate a vibration state score based on the vibration feature values ​​using the second preset dynamic scoring rule;

[0209] Step 103 is the core of the quantitative assessment of the wind turbine's mechanical condition. Its innovation lies in focusing the complex broadband vibration energy onto the fault-sensitive frequency band. By extracting vibration characteristic values ​​(F-values) and applying dynamic scoring rules, it achieves the following:

[0210] 1. Single-frequency vibration intensity quantification: accurately captures characteristic frequency energy concentration phenomena caused by bearing damage, impeller imbalance, etc.

[0211] 2. Early fault warning: When the total vibration is within the standard, potential faults are identified by the prominence of the harmonic components;

[0212] 3. Dynamic generation of status scores: The score is adaptively output based on the severity of the fault, avoiding the risk of missed detection by the traditional threshold method.

[0213] Optionally, see Figure 3 It can be seen that step 103, extracting vibration feature values ​​characterizing the intensity of single-frequency vibration from the vibration time-domain signal, includes:

[0214] Step 1031: Perform frequency domain transformation on the vibration time domain signal to obtain the vibration spectrum;

[0215] Step 1032: Extract the maximum vibration peak value of the vibration spectrum;

[0216] Step 1033: Calculate the effective vibration value for the entire time period based on the vibration time-domain signal;

[0217] Step 1034: Use the ratio of the maximum vibration peak value to the effective vibration value as the vibration characteristic value.

[0218] Step 1031: Obtain the vibration spectrum by frequency domain transformation;

[0219] Technical objective: To reveal the frequency distribution law of vibration energy and separate the background vibration from the fault characteristic frequency;

[0220] Conversion method:

[0221] Basic solution: Fast Fourier Transform (FFT), with a frequency range covering 0-5kHz (meeting the highest characteristic frequency requirement of the wind turbine);

[0222] Enhancement solution: Linear frequency modulation Z-transform (CZT) improves resolution in specific frequency bands, suitable for bearing fault detection;

[0223] Example of parameter requirements:

[0224] Frequency resolution ≤2Hz (precisely locates narrowband components);

[0225] Dynamic range ≥70dB (compatible with weak fault signals);

[0226] Example: For a 1500RPM motor driving a fan, with a motor base frequency f0=25Hz, using CZT focusing on the 20-30Hz frequency band, the resolution can reach 0.1Hz, accurately locating the 25.3Hz frequency component;

[0227] Step 1032: Extract the maximum vibration peak value;

[0228] Technical objective: To identify the amplitude spikes most sensitive to mechanical failures;

[0229] Extraction strategies (including optional options):

[0230] 1. Global peak extraction: Directly extract the point with the largest amplitude in the spectrum (suitable for significant faults);

[0231] 2. Peak value extraction of characteristic frequency bands (preferred):

[0232] Define the characteristic frequency range:

[0233] Lower limit: Rated motor speed × 2 / 60;

[0234] Upper limit: Motor rated speed × 5 / 60;

[0235] Example: 1500RPM fan → characteristic frequency range 50-125Hz;

[0236] Peak location method:

[0237] Option A: Scan for the maximum value within the defined range;

[0238] Option B: Identify harmonic clusters that are integer multiples of the fundamental frequency, and take the component with the largest amplitude.

[0239] Step 1033: Calculate the effective vibration value over the entire time period;

[0240] Technical objective: To characterize the total energy intensity of vibration signals;

[0241] Calculation specifications:

[0242] 1. Root mean square calculation in the time domain:

[0243] ;

[0244] 2. Unified units: The output of the accelerometer is directly used as the effective value without unit conversion;

[0245] Example of key requirements:

[0246] Sampling duration ≥ 10 fan rotation cycles (to ensure statistical representativeness);

[0247] Remove DC offset (hardware high-pass filtering or software zeroing).

[0248] Step 1034: Generate vibration characteristic value F;

[0249] definition:

[0250] (Dimensionless ratio);

[0251] Physical meaning:

[0252] Small F value (e.g., F<1.414) → Uniform vibration energy distribution → No significant single-frequency faults;

[0253] Large F value (e.g., F>3) → High concentration of energy at a specific frequency → Significant local fault exists;

[0254] Technical advantages: It is 5 times more sensitive to bearing damage than the traditional effective value of vibration velocity (ISO 10816 verification data).

[0255] The second preset dynamic scoring rule maps the F-value to a vibration state score:

[0256] ;

[0257] Scoring logic:

[0258] Segmentation design basis: Nonlinear relationship between the degree of failure of rotating machinery and F value (experimental calibration);

[0259] Project correspondence:

[0260] → Vibration status is normal;

[0261] → It is recommended to check the dynamic balance;

[0262] → The bearing is severely damaged and needs to be replaced;

[0263] The vibration state score generated in this step will be used in conjunction with the noise score and the age correction factor for decision-making.

[0264] Low noise score + high vibration score (e.g., S) noise =3, S v =7) → Bearing failure is the main cause → Triggering replacement instruction information;

[0265] Medium noise score + medium vibration score (e.g., S) noise =5, S v =5) → Slight impeller imbalance → Trigger dynamic balancing correction;

[0266] By focusing on characteristic frequencies and using ratio quantization, this scheme can distinguish:

[0267] Bearing outer ring damage: Energy concentration at the characteristic frequency at 3 times the rotational frequency (F>3);

[0268] Impeller fouling imbalance: prominent energy at the fundamental frequency (1 times the rotational frequency F≈2.5).

[0269] Optionally, step 1032, extracting the maximum vibration peak value of the vibration spectrum, includes: extracting the point with the largest amplitude value as the characteristic frequency peak value within the preset frequency range of the fan;

[0270] And / or,

[0271] Step 1033: Calculate the effective vibration value for the entire time period based on the vibration time-domain signal, including: calculating the root mean square value of acceleration for all sampling points of the vibration time-domain signal, and using the root mean square value of acceleration as the effective vibration value.

[0272] The core of step 1032 lies in locking down the sensitive frequency band of mechanical faults through a preset frequency range. The scientific basis for this is the strong correlation between the characteristic frequency of rotating machinery faults and the rotational speed (ISO 13373-2 standard). Compared with extracting the global maximum value, this method can eliminate irrelevant high-frequency noise interference and focus on specific harmonic vibrations caused by bearing damage, impeller imbalance, etc.

[0273] Engineering significance: The typical characteristic frequencies of wind turbine mechanical failures are concentrated within an integer multiple of the fundamental frequency of the rotational speed.

[0274] Lower limit setting: ≥2 times the fundamental frequency (excluding low-frequency noise such as electromagnetic interference from motors);

[0275] Upper limit setting: ≤5 times the base frequency (to avoid non-fault signals such as high-frequency resonance).

[0276] Methods for determining dynamic range:

[0277] 1. Basic solution: Fixed frequency range (e.g., 50-200Hz), suitable for constant speed fans;

[0278] 2. Preferred solution: Based on real-time dynamic speed calculation:

[0279] ;

[0280] in RPM is the rated speed of the motor.

[0281] Example: A 1500RPM fan:

[0282] ;

[0283] Special scene processing for peak extraction includes:

[0284] 1. Multi-peak competition:

[0285] When there are multiple peak values ​​with similar amplitudes within the preset range, frequency points that are integer multiples of the fundamental frequency are preferred.

[0286] Example: In the 50-125Hz range, if the peak value at 75Hz (3×25Hz) is 8.2g and the peak value at 100Hz (4×25Hz) is 8.0g, then the 75Hz component is taken.

[0287] 2. Sideband interference:

[0288] For gearbox fans, it is necessary to identify the modulation sidebands (such as meshing frequency ± rotational frequency) and take the maximum value in the sideband group;

[0289] Example: Meshing frequency 320Hz, sideband 318Hz / 322Hz, take the largest amplitude value among the three.

[0290] Step 1033 clearly defines the root mean square value of acceleration as the sole representation of the total energy of vibration, and its physical nature conforms to Newton's second law (F=ma), thus avoiding the conversion error introduced by the velocity / displacement dimensions.

[0291] The principle of calculating the root mean square value of acceleration:

[0292] ;

[0293] Note: a(n) is the time-domain sampled value of acceleration, in m / s²;

[0294] Signal preprocessing:

[0295] DC component elimination: hardware high-pass filter (cutoff frequency 0.5Hz) or software subtraction of the mean;

[0296] Outlier removal: If the amplitude of a sampling point exceeds 80% of the range, the resampling mechanism is triggered.

[0297] Connection with step 1031:

[0298] Frequency domain transformation (step 1031) and RMS value calculation (step 1033) require time window synchronization:

[0299] If the FFT uses a 4096-point Hanning window, then The calculation requires the same 4096 points of original signal;

[0300] Window function energy loss compensation coefficient: 1.5 for Hanning window and 2.23 for flat-top window.

[0301] Logical closed loop with step 1034:

[0302] The value of this project lies in:

[0303] From the preset fault frequency band → Characterize the intensity of local damage;

[0304] Time-domain computation → Characterizing overall vibrational energy;

[0305] F → Quantifies the proportion of fault energy, which is 3 times more sensitive than a single parameter.

[0306] Step 104: Calculate the service life correction factor using a piecewise increasing function based on the service life of the wind turbine;

[0307] Step 104 is the core module of the aging compensation in the wind turbine condition assessment system. Its innovation lies in quantifying the nonlinear degradation of equipment performance over time. Traditional methods ignore the impact of service life, leading to a misjudgment rate of approximately 30% when using the same standard for both new and old equipment. This solution achieves this through a piecewise increasing function:

[0308] 1. Performance degradation modeling: Reflects the irreversible aging process of wind turbine bearings, impeller fatigue, etc.

[0309] 2. Dynamic threshold adjustment: Avoids the dual problems of over-maintaining new equipment and under-maintaining old equipment.

[0310] The technical essence of piecewise increasing functions includes:

[0311] Nonlinear growth characteristics:

[0312] Initial use (0-2 years): Performance stable period, the age correction factor remains at the baseline value of 1 (excluding the effects of aging);

[0313] Mid-term use (3-5 years): slow decline period, with the annual correction factor increasing linearly (annual growth rate of approximately 0.05);

[0314] Late-term use (>5 years): Accelerated aging period, with an exponential increase in the age correction factor (annual growth rate ≥ 0.1).

[0315] Mathematical expression example:

[0316] ;

[0317] Application example: A certain fan has been in use for 6 years:

[0318] Interim revision: 1 + 0.05 × (5 - 2) = 1.15;

[0319] Later correction: 1.15 + 0.1 × (6 - 5) = 1.25;

[0320] This reflects that the growth rate in the later stage (0.1) is twice that of the growth rate in the middle stage (0.05).

[0321] Key points for implementation and examples of error control:

[0322] 1. How to obtain years of service:

[0323] Automatic recording: Cumulative operating hours of the wind turbine control system ÷ 8760 (annual hours);

[0324] Manual input: Maintenance personnel enter the equipment activation date;

[0325] 2. Function configurability:

[0326] For high-load scenarios (such as restaurant kitchens): the growth rate can be adjusted to 0.15 / year in the later stage;

[0327] For mild scenarios (such as home kitchens): the growth rate will remain at 0.08% per year in the later stages;

[0328] 3. Boundary handling:

[0329] When the number of years is less than or equal to 0, it must be set to 1 (to prevent input errors).

[0330] Limited to periods >15 years ≤3 (avoid over-correction).

[0331] Step 105: The noise status score, vibration status score, and age correction factor are weighted and fused to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

[0332] Step 105 is the multi-parameter integrated hub for condition assessment. It generates the final decision-making basis by fusing data from three dimensions: noise, vibration, and age. Its technical advantages are:

[0333] 1. Heterogeneous data normalization: mapping acoustic, mechanical, and aging parameters to a unified scoring system;

[0334] 2. Fault sensitivity optimization: Higher noise weight (human ears are more sensitive to abnormal noises);

[0335] 3. Precise maintenance decision-making: Output three levels of instruction information: "Good condition / Needs cleaning / Needs replacement".

[0336] The mathematical models for weighted fusion include:

[0337] Basic formula:

[0338] ;

[0339] Where:

[0340] : Noise state scoring weight (preferably 0.65);

[0341] : Vibration state scoring weight (preferably 0.35);

[0342] W1 + W2 = 1 and W1 > W2 (noise dominates the decision).

[0343] Parameter design basis:

[0344] The physical meaning of is the severity of abnormal noise, and the weight distribution is based on that the human ear's sensitivity to abnormal noise is 1.8 times that of vibration (ISO / TS 15666);

[0345] The physical meaning of is the intensity of mechanical failure, and the weight distribution is based on that vibration needs to reach the tactile perception threshold to significantly affect the experience;

[0346] The physical meaning of is the performance attenuation coefficient, and the weight distribution is based on experimental data showing that the misjudgment risk of old equipment is 40% higher.

[0347] The decision logic of the comprehensive state score includes:

[0348] Judgment rule:

[0349] 0 < Z ≤ 4 → Good state (no operation required);

[0350] 4 < Z ≤ 8 → Need to clean (dirt accumulation on impeller / air duct);

[0351] Z > 8 → Need to replace the machine (bearing / motor severely damaged).

[0352] Threshold design principle:

[0353] Cleaning threshold 4: Corresponding to the working condition of noise score 5 + vibration score 3 (about 60% performance attenuation);

[0354] Replacement threshold 8: Corresponding to the superposition effect of noise score 7 + vibration score 7 (> 80% performance attenuation).

[0355] Decision example:

[0356] Measured for a 3-year-old fan:

[0357] (moderate abnormal noise), S (slight vibration), ;

[0358] Overall score: Z = 1.05 × (0.65 × 5.2 + 0.35 × 3.8) ≈ 4.6 → Triggering the "Clean Needed" instruction message;

[0359] The technical collaboration and fault diagnosis logic includes:

[0360] 1. Multi-parameter cross-validation mechanism:

[0361] High noise + low vibration: Indicates impeller fouling (abnormal noise caused by airflow disturbance);

[0362] Low noise + high vibration: Indicates early bearing damage (vibration does not generate abnormal noise);

[0363] High parameters + long service life: Equipment is confirmed to be scrapped (comprehensive aging is irreversible);

[0364] 2. The amplification effect of age adjustment:

[0365] New equipment (2 years): Z new =1.0×(0.65×6.0+0.35×5.0)=5.65→ Needs cleaning;

[0366] Old equipment (7 years): Z old =1.35×(0.65×6.0+0.35×5.0)≈7.63→ Approaching the replacement threshold;

[0367] Under the same operating conditions, decisions regarding older equipment are more stringent.

[0368] Optionally, the weighted fusion satisfies at least one of the following:

[0369] Firstly, the first contribution weight of the noise state score is greater than the second contribution weight of the vibration state score;

[0370] The core technology behind this optional condition lies in prioritizing noise parameters in decision-making. Its scientific basis stems from the high sensitivity of the human ear to abnormal fan noise (psychoacoustic principles). Compared to vibration signals, abnormal noise indicates impeller damage, bearing dry friction, and other faults earlier, and directly impacts user experience.

[0371] Weighting relationship definition:

[0372] The first contribution weight W1 (noise) and the second contribution weight W2 (vibration) must satisfy W1 > W2. Typical implementations include:

[0373] Baseline ratios: W1=0.65, W2=0.35 (experimental verification of the optimal solution);

[0374] Extended range: W1 / W2 ratio between 1.5 and 2.5 (covering the sensitivity requirements of different fan types).

[0375] Secondly, the service life correction factor increases monotonically with the service life.

[0376] This condition requires that the service life correction factor change strictly and non-decreasingly with the service life, reflecting the irreversible degradation law of equipment performance.

[0377] Monotonicity guarantee method:

[0378] 1. Piecewise linear functions:

[0379] ;

[0380] Where k1>0, k2≥k1 (the growth rate in the later stage is not lower than that in the earlier stage).

[0381] 2. Exponential growth model:

[0382] ;

[0383] Necessity of the project:

[0384] Experimental data show that the rate of performance degradation of wind turbines increases with age:

[0385] 0-2 years: Annual decay rate <2% → Age correction factor ≈ 1;

[0386] 5-8 years: Annual degradation rate 8%~15% → Accelerated correction required;

[0387] Application example: In the 7th year of a certain wind turbine, if a linear model is used with k1=0.05 and k2=0.1, then... F; if an exponential model is adopted ,but .

[0388] Third, the output value of the weighted fusion is the product of the age correction factor and the weighted score, where the weighted score includes: the product of the noise state score and the first contribution weight, and the product of the vibration state score and the second contribution weight.

[0389] This condition explicitly defines the mathematical form of weighted fusion as a product relationship, ensuring that the years of service adjustment affects the overall score.

[0390] Basic formula:

[0391] ;

[0392] Structural advantages:

[0393] 1. Explainability: As an independent multiplier, it highlights the impact of aging on the overall condition;

[0394] 2. Fault isolation: When (For new equipment) Z is entirely determined by the real-time state;

[0395] 3. Sensitivity optimization: The age correction does not change the weight ratio of noise / vibration.

[0396] Example 2

[0397] Corresponding to the aforementioned embodiments of the method for detecting the operating status of wind turbines, this disclosure also provides embodiments of a system for detecting the operating status of wind turbines.

[0398] Figure 4 A schematic diagram of a wind turbine operating status detection system provided as an exemplary embodiment of this disclosure, the system comprising:

[0399] This disclosure provides a system for detecting the operating status of a wind turbine, the system comprising:

[0400] The acquisition module 21 is used to acquire the noise time-domain signal and vibration time-domain signal of the fan when the fan is running at its maximum speed.

[0401] The first scoring module 22 is used to extract acoustic feature values ​​representing the prominence of abnormal sounds from the noise time-domain signal, and generate a noise state score based on the acoustic feature values ​​through a first preset dynamic scoring rule.

[0402] The second scoring module 23 is used to extract vibration feature values ​​representing the intensity of single-frequency vibration from the vibration time-domain signal, and generate a vibration state score based on the vibration feature values ​​through a second preset dynamic scoring rule.

[0403] The service life correction module 24 is used to calculate the service life correction factor based on the service life of the wind turbine using a piecewise increasing function.

[0404] The third scoring module 25 is used to weight and fuse the noise status score, vibration status score and age correction factor to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

[0405] Optionally, the first scoring module 22 is specifically used for:

[0406] The sound pressure spectrum is obtained by performing frequency domain transformation on the noise time-domain signal;

[0407] Extract the peak value of the maximum sound pressure level in the sound pressure spectrum;

[0408] Calculate the effective sound pressure level over the entire time period based on the noise time-domain signal;

[0409] The difference between the effective sound pressure level and the peak value of the maximum sound pressure level is used as the acoustic characteristic value.

[0410] Optionally, the first scoring module 22 is specifically used to: identify abnormal frequency bands in the sound pressure spectrum where the gradient change rate is greater than or equal to the change rate threshold, and extract the peak value of the maximum sound pressure level from the abnormal frequency bands;

[0411] And / or,

[0412] The first scoring module 22 is specifically used to: calculate the root mean square value of sound pressure for all sampling points of the noise time-domain signal, and convert the root mean square value of sound pressure into a decibel value as the effective sound pressure level.

[0413] Optionally, the second scoring module 23 is specifically used for:

[0414] The vibration time-domain signal is converted to the frequency domain to obtain the vibration spectrum;

[0415] Extract the maximum vibration peak value from the vibration spectrum;

[0416] Calculate the effective vibration value over the entire time period based on the vibration time-domain signal;

[0417] The ratio of the maximum vibration peak value to the effective vibration value is used as the vibration characteristic value.

[0418] Optionally, the second scoring module 23 is specifically used to: extract the maximum amplitude point as the characteristic frequency peak within the preset frequency range of the wind turbine;

[0419] And / or,

[0420] The second scoring module 23 is specifically used to calculate the root mean square value of acceleration for all sampling points of the vibration time-domain signal, and the root mean square value of acceleration is used as the effective vibration value.

[0421] Optionally, the weighted fusion satisfies at least one of the following:

[0422] The first contribution weight of the noise state score is greater than the second contribution weight of the vibration state score;

[0423] The service life correction factor increases monotonically with the service life.

[0424] The output value of the weighted fusion is the product of the age correction factor and the weighted score, where the weighted score includes: the product of the noise state score and the first contribution weight, and the product of the vibration state score and the second contribution weight.

[0425] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure solution according to actual needs.

[0426] Example 3

[0427] Figure 5 This is a schematic diagram of the structure of a smart appliance according to an example embodiment of the present disclosure. The smart appliance includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the fan operation status detection method described in any of the above embodiments. Figure 5 The smart appliance 90 shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments disclosed herein.

[0428] like Figure 5 As shown, the intelligent appliance 90 can be represented in the form of a general-purpose computing device, such as a server device. The components of the intelligent appliance 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different system components (including memory 92 and processor 91).

[0429] Bus 93 includes a data bus, an address bus, and a control bus.

[0430] The memory 92 may include volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0431] The memory 92 may also include a program tool 925 (or utility) having a set (at least one) program module 924, such program module 924 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0432] The processor 91 executes various functional applications and data processing by running computer programs stored in the memory 92, such as the method for detecting the operating status of a wind turbine provided in any of the above embodiments.

[0433] The smart appliance 90 can also communicate with one or more external devices 94 (e.g., keyboards, pointing devices, etc.). This communication can be achieved through input / output (I / O) interface 95. Furthermore, the smart appliance 90 can also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 96. As shown in the figure, network adapter 96 communicates with other modules of the smart appliance 90 via bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the smart appliance 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0434] The smart appliance can be controlled by a voice module, which is equipped with a controller, a voice receiving module, and a voice parsing module. The voice receiving module receives user commands, and the voice parsing module parses the commands. Based on the parsed commands, the controller controls the smart appliance to perform corresponding operations, thereby realizing intelligent control of the smart appliance and improving the user experience.

[0435] It should be noted that although several units / modules or sub-units / modules of smart appliances have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0436] Example 4

[0437] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for detecting the operating status of a wind turbine provided in any of the above embodiments.

[0438] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0439] Example 5

[0440] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the method for detecting the operating status of a wind turbine as described above.

[0441] The program code for executing the computer program product of this disclosure can be written in any combination of one or more programming languages, and the program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0442] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A method for detecting the operating status of a wind turbine, characterized in that, The detection method includes: With the fan operating at its maximum speed, the time-domain noise signal and the time-domain vibration signal of the fan are collected. Acoustic feature values ​​representing the prominence of abnormal sounds are extracted from the noise time-domain signal, and a noise state score is generated based on the acoustic feature values ​​through a first preset dynamic scoring rule; Vibration feature values ​​characterizing the intensity of single-frequency vibration are extracted from the vibration time-domain signal, and a vibration state score is generated based on the vibration feature values ​​through a second preset dynamic scoring rule; Based on the service life of the wind turbine, the service life correction factor is calculated using a piecewise increasing function. The noise status score, the vibration status score, and the age correction factor are weighted and fused to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

2. The detection method according to claim 1, characterized in that, The extraction of acoustic feature values ​​characterizing the prominence of abnormal sounds from the time-domain noise signal includes: The noise time-domain signal is converted into a frequency domain to obtain the sound pressure spectrum; Extract the maximum sound pressure level peak value of the sound pressure spectrum; Calculate the effective sound pressure level for the entire time period based on the noise time-domain signal; The difference between the effective sound pressure level and the peak value of the maximum sound pressure level is used as the acoustic characteristic value.

3. The detection method according to claim 2, characterized in that, The step of extracting the maximum sound pressure level peak value of the sound pressure spectrum includes: identifying abnormal frequency bands in the sound pressure spectrum where the gradient change rate is greater than or equal to a change rate threshold, and extracting the maximum sound pressure level peak value from the abnormal frequency bands; And / or, The step of calculating the effective sound pressure level over the entire time period based on the noise time-domain signal includes: calculating the root mean square value of the sound pressure for all sampling points of the noise time-domain signal, and converting the root mean square value of the sound pressure into a decibel value as the effective sound pressure level.

4. The detection method according to claim 1, characterized in that, The extraction of vibration feature values ​​characterizing the intensity of single-frequency vibration from the vibration time-domain signal includes: The vibration time-domain signal is converted into a frequency domain to obtain the vibration spectrum; Extract the maximum vibration peak value of the vibration spectrum; Calculate the effective vibration value for the entire time period based on the vibration time-domain signal; The ratio of the maximum vibration peak value to the effective vibration value is used as the vibration characteristic value.

5. The detection method according to claim 4, characterized in that, The step of extracting the maximum vibration peak value of the vibration spectrum includes: extracting the point with the largest amplitude value as the characteristic frequency peak value within the preset frequency range of the fan; And / or, The calculation of the effective vibration value for the entire time period based on the vibration time-domain signal includes: calculating the root mean square value of acceleration for all sampling points of the vibration time-domain signal, and using the root mean square value of acceleration as the effective vibration value.

6. The detection method according to claim 1, characterized in that, The weighted fusion satisfies at least one of the following: The first contribution weight of the noise state score is greater than the second contribution weight of the vibration state score; The age correction factor increases monotonically with the service life. The output value of the weighted fusion is the product of the age correction factor and the weighted score, wherein the weighted score includes: the product of the noise state score and the first contribution weight, and the product of the vibration state score and the second contribution weight.

7. A system for detecting the operating status of a wind turbine, characterized in that, The detection system includes: The acquisition module is used to acquire the noise time-domain signal and vibration time-domain signal of the fan when the fan is running at its maximum speed. The first scoring module is used to extract acoustic feature values ​​representing the prominence of abnormal sounds from the noise time-domain signal, and generate a noise state score based on the acoustic feature values ​​through a first preset dynamic scoring rule. The second scoring module is used to extract vibration feature values ​​characterizing the intensity of single-frequency vibration from the vibration time-domain signal, and generate a vibration state score based on the vibration feature values ​​through a second preset dynamic scoring rule. The service life correction module is used to calculate the service life correction factor based on the service life of the wind turbine using a piecewise increasing function. The third scoring module is used to weight and fuse the noise status score, the vibration status score, and the age correction factor to generate a comprehensive status score; the comprehensive status score is used to determine the operating status of the wind turbine.

8. A smart electrical appliance, comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the method for detecting the operating status of the wind turbine as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for detecting the operating status of the wind turbine as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for detecting the operating status of the wind turbine as described in any one of claims 1 to 6.

Images