Damage assessment method and system for key component of low-speed heavy-load equipment and electronic equipment

Abstract

The invention relates to the technical field of equipment fault diagnosis, in particular to a damage assessment method and system for key components of low-speed heavy-load equipment and electronic equipment. The method comprises the following steps: acquiring and processing an acoustic emission event signal of a key component to obtain a time-frequency domain signal; dividing a plurality of frequency bands, calculating the energy ratio of each frequency band, screening effective signals in a fault state based on a peak amplitude threshold value and a high-frequency band energy ratio threshold value, and calculating a weighted impact number, energy and count; substituting into a logarithmic linear judgment model, and obtaining a corrected comprehensive index in combination with the total dominant frequency band type and the high-frequency energy ratio; threshold values are divided according to the health state data, and the fault level is judged through the corrected comprehensive indexes. According to the method, weak fault signals are efficiently extracted, a quantitative evaluation system is constructed, multiple scenes are adapted, the problems of difficult signal extraction, classification missing and insufficient adaptability under the low-speed and heavy-load working condition are solved, and accurate damage evaluation and early warning are achieved.

Description

Technical Field

[0001] This invention relates to the field of equipment fault diagnosis technology, and in particular to a damage assessment method, system and electronic equipment for key components of low-speed heavy-load equipment. Background Technology

[0002] Low-speed, heavy-duty equipment is widely used in key industrial sectors such as energy, machinery, and chemicals. Its critical components, such as bearings, gears, and spindles, directly determine the equipment's operational stability and reliability. This type of equipment operates under harsh conditions of extremely low speeds and high impact loads for extended periods. Critical components are prone to progressive damage due to friction, loosening, crack propagation, and wear. Failure to promptly and accurately identify the damage and assess its severity can easily lead to sudden component failure, resulting in equipment downtime, production interruptions, or even safety accidents, causing significant economic losses and safety risks.

[0003] Currently, the monitoring of key components in low-speed, heavy-load equipment mainly relies on traditional vibration monitoring and acoustic emission monitoring technologies. Traditional vibration monitoring, under extremely low speed and high impact load conditions, is susceptible to low-frequency background interference, making it difficult to extract weak fault signals and accurately quantify the degree of damage. While acoustic emission monitoring technology is sensitive to transient faults, existing solutions have significant shortcomings: some focus solely on optimizing sensor hardware to improve sensitivity, failing to address the core issue of separating weak signals from interference through signal characteristic analysis; some technologies use only single-dimensional characteristic parameters for fault diagnosis, lacking the utilization of multi-dimensional comprehensive characteristics such as acoustic emission signal intensity, energy, frequency, and frequency distribution, and also lack comprehensive quantitative assessment methods, making it difficult to distinguish damage levels; they also exhibit low sensitivity to early-stage high-frequency crack-type faults and lack flexible adjustment mechanisms adapted to different equipment and operating conditions.

[0004] Therefore, there is an urgent need to develop a damage assessment method that can adapt to low-speed, heavy-load, and harsh operating conditions, effectively extract weak fault signals, accurately classify and assess damage, and adapt to different equipment. Summary of the Invention

[0005] The purpose of this invention is to provide a damage assessment method, system, and electronic device for key components of low-speed, heavy-load equipment, in order to solve the problems of difficulty in extracting weak fault signals, lack of quantitative damage classification, and insufficient robustness of solutions under low-speed, heavy-load conditions in the prior art.

[0006] The technical solution of this invention is: a damage assessment method for key components of low-speed heavy-load equipment, comprising: The system collects acoustic emission signals from key components of low-speed heavy-load equipment during operation, filters out noise to obtain time-domain signals, and performs frequency-domain transformation on the time-domain signals to obtain frequency-domain signals. The frequency domain signal is divided into corresponding frequency bands according to the preset frequency band, the energy proportion of each frequency band is calculated, and the preset peak amplitude threshold and high frequency band energy proportion threshold are used as fault feature judgment conditions to judge the fault features. The time-domain and frequency-domain signals of acoustic emission events in fault states are selected as valid signals. Based on each group of valid signals, the corresponding amplitude weight and frequency band weight are obtained. The weighted impact number, weighted energy, and weighted count are calculated by combining the amplitude weight and frequency band weight of each group. The weighted impact number, weighted energy, weighted count, and overall dominant frequency band type are substituted into the logarithmic linear judgment model to calculate the comprehensive index. The fault level threshold is divided based on the comprehensive index under the healthy state. The comprehensive index under the fault state is linearly corrected by combining the overall high-frequency energy ratio, and the fault level is judged by the corrected comprehensive index.

[0007] Preferably, the method for obtaining the corresponding amplitude weight and frequency band weight based on each group of valid signals includes: The ratio of the peak amplitude of each valid signal group to the preset healthy amplitude is calculated as the amplitude weight of each group; the preset healthy amplitude is obtained based on the maximum allowable safe amplitude of the component. The frequency corresponding to the maximum spectral energy of each group of effective signals is taken as the main frequency, and the frequency band weight of each group is determined according to the frequency band where the main frequency is located.

[0008] Preferably, the method for determining the frequency band weight of each group based on the frequency band where the main frequency is located is as follows: if the main frequency belongs to the dominant frequency band of the effective signal in the group, the frequency band weight is set to 1; if the main frequency does not belong to the dominant frequency band, the frequency band weight is a weight attenuation factor; the weight attenuation factor is a preset value; the dominant frequency band of the effective signal in the group refers to the frequency band with the largest energy proportion in the effective signal in the group.

[0009] Preferably, the method for calculating the weighted impact number, weighted energy, and weighted count by combining the amplitude weight and frequency band weight of each group includes: There are n groups of valid signals. The weighted collision number is obtained by summing the products of the amplitude weight and the frequency band weight corresponding to the n groups of valid signals. The weighted energy is obtained by multiplying the energy of n groups of valid signals with their corresponding frequency band weights and summing the results. The frequency band weights corresponding to the n groups of valid signals are accumulated to obtain a weighted count.

[0010] Preferably, the method for calculating the comprehensive index is as follows: the comprehensive index is obtained by summing the product of the overall dominant frequency band type code and the frequency band type correction coefficient, the product of the logarithm of the weighted impact number and the impact occurrence frequency weight coefficient, the product of the logarithm of the weighted energy and the damage energy weight coefficient, and the product of the logarithm of the weighted count and the activity level weight coefficient. The sum of the impact frequency weighting coefficient, the damage energy weighting coefficient, and the activity level weighting coefficient is 1.

[0011] Preferably, the method for determining the overall dominant frequency band type code is as follows: The frequency band energy percentage of all valid signals is statistically analyzed to obtain the overall low-frequency energy percentage, overall mid-frequency energy percentage, and overall high-frequency energy percentage. The frequency band corresponding to the largest frequency band energy percentage is taken as the overall dominant frequency band. When the overall dominant frequency band is a low frequency band, the corresponding overall dominant frequency band type code is 0; when the overall dominant frequency band is a mid frequency band, the corresponding overall dominant frequency band type code is 1; and when the overall dominant frequency band is a high frequency band, the corresponding overall dominant frequency band type code is 2.

[0012] Preferably, the method for classifying fault level thresholds based on comprehensive indicators of health status is as follows: A comprehensive index set of acoustic emission events under all healthy conditions is calculated, and the mean and standard deviation of the comprehensive index set are taken. Two comprehensive index thresholds are set based on the mean and standard deviation. The comprehensive index thresholds are used to classify the fault levels, which include minor abnormalities, progressive damage, and severe damage.

[0013] Preferably, the method for linearly correcting the comprehensive index under fault conditions by combining the overall high-frequency energy ratio is as follows: set a correction gain coefficient, weight the overall high-frequency energy ratio with the correction gain coefficient, add the weighted overall high-frequency energy ratio to 1 to obtain the correction coefficient; multiply the correction coefficient by the comprehensive index to obtain the corrected comprehensive index.

[0014] On the other hand, this application also discloses a damage assessment system for key components of low-speed heavy-load equipment, including: The signal processing module is used to collect signals of acoustic emission events during the operation of key components of low-speed heavy-load equipment, filter noise from the signals to obtain time-domain signals, and perform frequency-domain transformation on the time-domain signals to obtain frequency-domain signals. The preliminary judgment module is used to divide the frequency domain signal into corresponding frequency bands according to the preset frequency band, calculate the energy proportion of each frequency band, and use the preset peak amplitude threshold and high frequency band energy proportion threshold as fault feature judgment conditions to judge the fault features. The feature parameter module is used to filter the time-domain and frequency-domain signals of acoustic emission events in fault states as valid signals. Based on each group of valid signals, the corresponding amplitude weight and frequency band weight are obtained. The weighted impact number, weighted energy and weighted count are calculated by combining the amplitude weight and frequency band weight of each group. The fault level judgment module is used to substitute the weighted impact number, weighted energy, weighted count, and overall dominant frequency band type into the logarithmic linear judgment model to calculate the comprehensive index. Based on the comprehensive index under the healthy state, the fault level threshold is divided. The comprehensive index under the fault state is linearly corrected by combining the overall high-frequency energy ratio, and the fault level is judged by the corrected comprehensive index.

[0015] On the other hand, this application further discloses an electronic device, including: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a damage assessment method for key components of low-speed heavy-load equipment as described in any one of the above.

[0016] Compared with the prior art, the advantages of the present invention are: (1) This invention constructs a full-process signal enhancement scheme to efficiently solve the problem of weak fault signal extraction. Through the design of noise filtering, frequency band matching, weight enhancement and parameter accumulation, noise is first filtered to obtain effective signals, frequency bands are divided according to fault characteristics and the dominant frequency band is locked, and the quantization weight is calculated by combining the peak amplitude and the healthy amplitude. Then, through the accumulation operation of weighted impact number, energy and count, the weak fault signal features are accumulated, and even if the signal is covered by low frequency interference, it can still be highlighted. Compared with the single feature extraction method, it solves the pain point of difficult signal capture in low speed heavy load scenario from the root.

[0017] (2) This invention establishes a scientific quantitative assessment system to achieve damage grading and early warning. It constructs an assessment logic of comprehensive index quantification, health threshold grading and high-frequency correction early warning: through weighted parameter logarithmic transformation and dominant frequency band coding correction, a quantitative index that linearly reflects the damage evolution is obtained; based on the mean and standard deviation of healthy samples, a threshold is set to clarify the three-level classification of minor abnormalities, developing damage and severe damage; the overall high-frequency energy ratio correction is introduced to amplify the early crack signal, and an early warning can be given if the conventional threshold is not reached, thereby improving the sensitivity of high-risk fault identification and providing accurate quantitative basis for operation and maintenance.

[0018] (3) The present invention has outstanding adaptability and robustness, and is suitable for industrial applications in multiple scenarios. The core parameters adopt a preset benchmark and calibrable design. The preset frequency band, health amplitude, and various weight coefficients can be adjusted as needed to adapt to different equipment and key components. Through logarithmic transformation, dominant frequency band filtering, and dual correction, the influence of parameter magnitude fluctuation and operating condition interference is reduced, and stable evaluation accuracy is maintained under complex operating conditions such as speed and load fluctuation, which significantly improves industrial applicability. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of a damage assessment method for key components of low-speed heavy-load equipment as described in this invention. Figure 2 This is a structural block diagram of a damage assessment system for key components of low-speed heavy-load equipment according to the present invention. Detailed Implementation

[0020] The present invention will be further described in detail below with reference to specific embodiments: This application applies to damage monitoring of critical components such as wind turbine main shafts, heavy-duty gearboxes, and large compressor main shafts in the energy and machinery industries. These devices operate under low-speed, high-impact load conditions for extended periods, making fault signals susceptible to low-frequency interference. This invention addresses the core problems of existing technologies—difficulty in extracting weak fault signals, lack of quantitative damage grading, and insufficient robustness to adapt to different equipment and operating conditions—through a technical solution involving acoustic emission signal acquisition, frequency band division, multi-dimensional weighting, and comprehensive index evaluation. This enables accurate damage assessment and early warning.

[0021] like Figure 1 As shown, a damage assessment method for key components of low-speed heavy-load equipment includes: S1. Collect the acoustic emission event signals of key components of low-speed heavy-load equipment during operation, filter the noise in the signal to obtain the time domain signal, and perform frequency domain transformation on the time domain signal to obtain the frequency domain signal.

[0022] Specifically, low-speed heavy-duty equipment refers to mechanical equipment with extremely low speed and large impact load, such as heavy-duty gearboxes, compressors and wind turbines. Key components of this type of equipment, such as bearings, gears and main shafts, are prone to cracks, wear and pitting during operation, so it is necessary to detect and resolve faults in a timely manner.

[0023] Acoustic emission events refer to detectable acoustic emission signals with specific signal characteristics generated by key components of low-speed, heavy-load equipment during fault evolution or energy release. They are the core analytical object for quantitative assessment of acoustic emission damage. The process involves acquiring acoustic emission event signals and filtering out noise. The filtered time-domain signal is then transformed in the frequency domain to obtain the frequency domain signal. The frequency domain transformation includes short-time Fourier transform and continuous wavelet transform.

[0024] In one implementation, for monitoring the wind turbine main shaft bearing, an acoustic emission sensor is installed at the bearing end cover. The noise threshold is calibrated by collecting ambient noise when the equipment is stopped, ensuring that low-frequency interference caused by nacelle vibration is filtered out. The filtered acoustic emission signal is converted into a frequency domain signal by continuous wavelet transform.

[0025] S2. Divide the frequency domain signal into corresponding frequency bands according to the preset frequency band, calculate the energy proportion of each frequency band, and use the preset peak amplitude threshold and high frequency band energy proportion threshold as fault feature judgment conditions to perform fault feature judgment.

[0026] Specifically, multiple frequency bands are preset, that is, multiple different frequency ranges, to divide the frequency domain signal into multiple frequency bands; the energy of each frequency band refers to the sum of the energy of all frequency points extracted within the corresponding frequency band, that is, the cumulative energy of the frequency band; the energy ratio is the ratio of the energy of the corresponding frequency band to the total energy.

[0027] In one implementation, two frequency thresholds are set: 100kHz and 300kHz. The low-frequency band has frequencies less than 100kHz, the mid-frequency band has frequencies greater than or equal to 100kHz and less than 300kHz, and the high-frequency band has frequencies greater than or equal to 300kHz. The energy percentage is calculated by measuring the ratio of energy in each frequency band to the total energy, and the frequency band with the largest energy percentage is selected as the dominant frequency band. Simultaneously, a peak amplitude threshold and a high-frequency band energy percentage threshold are preset. If the peak amplitude of a certain acoustic emission event exceeds the peak amplitude threshold, or the high-frequency band energy percentage exceeds the high-frequency band energy percentage threshold, a fault is identified, and the component is in a faulty state; otherwise, it is in a healthy state.

[0028] S3. Select the time-domain and frequency-domain signals of acoustic emission events in the fault state as valid signals. Based on each group of valid signals, obtain the corresponding amplitude weight and frequency band weight. Combine the amplitude weight and frequency band weight of each group to calculate the weighted impact number, weighted energy and weighted count.

[0029] Specifically, the time-domain and frequency-domain signals corresponding to the acoustic emission events identified as faulty in step S2 are filtered as valid signals to ensure that subsequent analysis focuses on fault-related signals. Each group of valid signals is analyzed to obtain the amplitude weight and frequency band weight for each group. The amplitude weight represents the ratio of the peak amplitude of the current signal to the amplitude in a healthy state. The frequency band weight is a weight parameter used for calculating weighted parameters, set based on the frequency band division and the frequency band type to which the dominant frequency of the acoustic emission event belongs. Through preset weighted calculation rules, the weighted impact number, weighted energy, and weighted count are obtained, respectively reflecting the comprehensive characteristics of the intensity and frequency of the fault event, the energy characteristics of the accumulated fault, and the activity level of the fault event.

[0030] S4. Substitute the weighted impact number, weighted energy, weighted count, and overall dominant frequency band type into the logarithmic linear judgment model to calculate the comprehensive index. Based on the comprehensive index under the healthy state, classify the fault level threshold. Combine the overall high-frequency energy ratio to linearly correct the comprehensive index under the fault state, and use the corrected comprehensive index to judge the fault level.

[0031] Specifically, the dominant frequency band type is determined. The weighted impact number, weighted energy, and weighted count obtained in step S3 are substituted into the logarithmic linear decision model. Simultaneously, the dominant frequency band type is introduced to correct the model calculation process, ultimately obtaining a comprehensive index that reflects the fault state. This index integrates multi-dimensional information such as fault intensity, energy, frequency, and characteristic frequency bands. Comprehensive index data of multiple sets of acoustic emission events under the healthy state of key components are collected to construct a comprehensive health state index set. Based on this set, two fault level thresholds are defined to distinguish damage of different severity.

[0032] The overall high-frequency energy ratio of all valid signals is calculated. Through a preset linear correction logic, the overall high-frequency energy ratio is used to correct the comprehensive index under fault conditions. Finally, the fault level of the key component is determined by comparing the corrected comprehensive index with the fault level threshold.

[0033] In one implementation, taking wind turbine main shaft bearing monitoring as an example, the overall high-frequency energy proportion of all valid signals is statistically analyzed to determine the overall dominant frequency band as the high-frequency band. After correction by incorporating it into a log-linear judgment model, a comprehensive index is obtained. 100 sets of healthy acoustic emission event data are collected from the wind turbine main shaft bearing under brand-new conditions, covering signals under different shaft speeds and load conditions. A comprehensive health status index set is calculated, and a first threshold and a second threshold are determined based on the mean and standard deviation of this set. The first threshold is the boundary between minor abnormalities and progressive damage, and the second threshold is the boundary between progressive damage and severe damage.

[0034] Further calculate the overall high-frequency energy ratio of all current valid signals, and correct the initial comprehensive index according to the linear correction rule. If the corrected comprehensive index is less than the first threshold, it is judged as a minor anomaly, such as slight wear of the rolling elements. The operation and maintenance personnel can continuously monitor it in conjunction with the wind turbine operation scheduling plan. If the corrected comprehensive index is greater than or equal to the first threshold and less than the second threshold, it is judged as developing damage, such as the early stage of crack propagation. The wind turbine needs to be scheduled for a shutdown window for bearing inspection or maintenance. If it is greater than or equal to the second threshold, it is judged as serious damage, such as crack penetration or rolling element peeling. The turbine needs to be shut down immediately to replace the bearing to avoid the fault from expanding and causing the wind turbine main shaft to jam, the nacelle to vibrate more, and thus causing the entire wind turbine to shut down or more serious equipment accidents.

[0035] In summary, this application, through a closed-loop design encompassing signal acquisition and processing, preliminary fault diagnosis, effective signal screening, characteristic parameter calculation, comprehensive index construction and correction, and fault classification, forms a complete technical solution for damage assessment adapted to low-speed, heavy-load operating conditions. This solution precisely addresses three core technical challenges in existing technologies: difficulty in extracting weak fault signals under low-speed, heavy-load conditions; lack of quantitative damage classification; and insufficient robustness in solution adaptation. Regarding weak fault signal extraction, a collaborative design employing noise filtering, frequency domain transformation, weighted quantization, and parameter accumulation is used to filter low-frequency interference, pinpoint fault characteristic frequency bands, and accumulate and strengthen weak signals for effective extraction. For the lack of quantitative damage classification, dual quantization of amplitude and frequency band is used, integrating multi-dimensional weighted parameters to construct a comprehensive index. Thresholds are set based on healthy samples to clearly distinguish three fault levels, achieving quantitative classification. Regarding insufficient robustness, core parameters support on-demand calibration to adapt to different equipment and components. Logarithmic transformation and a dual correction mechanism reduce the impact of operating condition interference and parameter magnitude fluctuations, ensuring stable and accurate assessment. This invention enables accurate extraction of weak fault signals, quantitative damage classification, and robust adaptation to multiple scenarios, providing a scientific and reliable solution for damage assessment of key components in low-speed heavy-load equipment.

[0036] Based on the aforementioned damage assessment method for key components of low-speed heavy-load equipment, to further elaborate on the solution of this application, this application also provides a method for obtaining corresponding amplitude weights and frequency band weights based on each group of valid signals, including: The ratio of the peak amplitude of each valid signal group to the preset healthy amplitude is calculated as the amplitude weight of each group; the preset healthy amplitude is obtained based on the maximum allowable safe amplitude of the component. The frequency corresponding to the maximum spectral energy of each group of effective signals is taken as the main frequency, and the frequency band weight of each group is determined according to the frequency band where the main frequency is located.

[0037] Specifically, the preset health amplitude is determined based on the maximum permissible safe amplitude of key components, serving as a reference benchmark for signal strength. For each group of valid signals, the amplitude weight of each group of valid signals is obtained through correlation calculation between its peak amplitude and the preset health amplitude, used to quantify the contribution of fault signals of different intensities. Simultaneously, in the frequency domain data of each group of valid signals, the frequency corresponding to the maximum spectral energy is located as the dominant frequency of the acoustic emission event, used to reflect the frequency with the most significant fault characteristics. Based on the frequency band to which the dominant frequency belongs, the frequency band weight of each group of valid signals is determined to enhance the signal contribution of the fault characteristic frequency band.

[0038] Methods for determining the band weight of each group based on the band where the dominant frequency is located include: If the main frequency belongs to the dominant frequency band of the group of valid signals, the frequency band weight is set to 1; if the main frequency does not belong to the dominant frequency band, the frequency band weight is the weight attenuation factor; where the weight attenuation factor is a preset value; the dominant frequency band of the group of valid signals refers to the frequency band with the largest energy proportion in the group of valid signals.

[0039] Specifically, the dominant frequency refers to the frequency corresponding to the maximum spectral energy in the group of effective signals. First, the low-frequency energy ratio, mid-frequency energy ratio, and high-frequency energy ratio of each group of effective signals are calculated, and the frequency band corresponding to the maximum energy ratio is taken as the dominant frequency band. It is determined whether the dominant frequency is within the dominant frequency band. If it is, the frequency band weight of the group of effective signals is set to 1; otherwise, the frequency band weight of the group of effective signals is equal to the weight attenuation factor. The weight attenuation factor has a value greater than 0 and less than 1. In some implementations, the value range is [0.3, 0.6].

[0040] The methods for calculating the weighted impact number, weighted energy, and weighted count by combining the amplitude weight and frequency band weight of each group include: There are n groups of valid signals. The weighted collision number is obtained by summing the products of the amplitude weight and the frequency band weight corresponding to the n groups of valid signals. The weighted energy is obtained by multiplying the energy of n groups of valid signals with their corresponding frequency band weights and summing the results. The frequency band weights corresponding to the n groups of valid signals are accumulated to obtain a weighted count.

[0041] Specifically, after noise filtering and fault feature screening of all acoustic emission events, the time-domain signal and frequency-domain signal corresponding to the acoustic emission events that are determined to be in a fault state are finally obtained. The two are used as a set of effective signals for subsequent damage quantification assessment and analysis.

[0042] In one implementation, there are n groups of valid signals, and the frequency band weight and amplitude weight of each group of valid signals have been determined. The weighted impact number is obtained according to the following formula. Weighted energy and weighted count : , , , in, The amplitude weights of the i-th group of valid signals are... The frequency band weights of the i-th group of valid signals are... Let be the energy of the i-th valid signal group.

[0043] The method for calculating the comprehensive index is as follows: the comprehensive index is obtained by summing the product of the overall dominant frequency band type code and the frequency band type correction coefficient, the product of the logarithm of the weighted impact number and the impact frequency weight coefficient, the product of the logarithm of the weighted energy and the damage energy weight coefficient, and the product of the logarithm of the weighted count and the activity level weight coefficient; wherein, the sum of the impact frequency weight coefficient, the damage energy weight coefficient and the activity level weight coefficient is 1.

[0044] The method for determining the overall dominant frequency band type coding is as follows: The frequency band energy percentage of all valid signals is statistically analyzed to obtain the overall low-frequency energy percentage, overall mid-frequency energy percentage, and overall high-frequency energy percentage. The frequency band corresponding to the largest frequency band energy percentage is taken as the overall dominant frequency band. When the overall dominant frequency band is a low frequency band, the corresponding overall dominant frequency band type code is 0; when the overall dominant frequency band is a mid frequency band, the corresponding overall dominant frequency band type code is 1; and when the overall dominant frequency band is a high frequency band, the corresponding overall dominant frequency band type code is 2.

[0045] Specifically, determining the overall dominant frequency band type involves statistically analyzing the energy and proportion of each frequency band of all valid signals. Since the threshold for dividing the frequency bands remains consistent throughout the implementation process, the frequency ranges of the low-frequency, mid-frequency, and high-frequency bands for each group of valid signals are consistent. The frequency band corresponding to the highest energy proportion is taken as the overall dominant frequency band, and its value is assigned to the overall dominant frequency band type code according to the aforementioned rules for determining the overall dominant frequency band type code.

[0046] The comprehensive index is calculated by substituting the overall dominant frequency band type code, weighted impact number, weighted energy, and weighted count into the log-linear decision model. The overall dominant frequency band type code is multiplied by the frequency band type correction coefficient, and then the logarithm of the weighted impact number, weighted energy, and weighted count is taken and multiplied by their respective weight coefficients. All the product results are summed to obtain the comprehensive index result.

[0047] In one implementation, the calculation formula for the log-linear criterion model is: , in, , and These are the impact frequency weighting coefficient, damage energy weighting coefficient, and activity level weighting coefficient, respectively. In this embodiment, their values ​​are 0.3, 0.5, and 0.2, respectively. This is the spectrum type correction factor, which is set to 0.3 in this embodiment. The overall dominant frequency band type is encoded, with values ​​including 0, 1, and 2, representing the low-frequency band, mid-frequency band, and high-frequency band as the overall dominant frequency band, respectively.

[0048] The method for classifying fault level thresholds based on comprehensive indicators of the health status of key components is as follows: Calculate a comprehensive index set for acoustic emission events under all healthy conditions, take the mean and standard deviation of the comprehensive index set, and set two comprehensive index thresholds based on the mean and standard deviation; the comprehensive index thresholds are used to classify the fault level, which includes minor abnormality, progressive damage and severe damage.

[0049] Specifically, by collecting multiple sets of acoustic emission event data on the health status of key components, covering signal samples under typical equipment operating conditions, a comprehensive health status index set is constructed. Statistical analysis is performed on all comprehensive index values ​​in this set to calculate the mean μ and standard deviation σ of the set. Based on the statistical results, two fault level thresholds are set: the first threshold... Second threshold .in, As a dividing line between minor abnormalities and developmental damage, As a dividing line between developmental injury and severe injury.

[0050] In one implementation, the first threshold Second threshold In other implementations, different calculation formulas for the first and second thresholds are set according to the actual working conditions or equipment.

[0051] The method for linearly correcting the comprehensive index under fault conditions by combining the overall high-frequency energy ratio is as follows: set a correction gain coefficient, weight the overall high-frequency energy ratio with the correction gain coefficient, add the weighted overall high-frequency energy ratio to 1 to obtain the correction coefficient, and multiply the correction coefficient by the comprehensive index to obtain the corrected comprehensive index.

[0052] Specifically, the overall high-frequency energy proportion directly reflects the proportion of high-frequency components in the fault signal. These high-frequency components are often directly related to more severe damage forms such as crack initiation, propagation, and fracture of critical components. Such damage releases more high-frequency acoustic emission energy during its evolution; a higher proportion indicates a more pronounced trend towards a dangerous state. Setting a correction gain coefficient essentially assigns a reasonable weight to the high-frequency energy proportion, strengthening the influence of high-frequency components on the overall index through weighted calculations, thus enabling the overall index to more accurately reflect the actual severity and evolutionary risk of the damage.

[0053] The comprehensive index proposed in this application is constructed based solely on weighted impact number, weighted energy, weighted count, and dominant frequency band type. While it can reflect the basic state of damage, it lacks sensitivity to early high-frequency crack-type faults. These faults may not initially trigger significant changes in weighted parameters, causing the comprehensive index to fail to reach the corresponding fault level threshold, resulting in a delayed early warning. By introducing a linear correction based on the overall high-frequency energy proportion, the contribution of early high-frequency crack signals can be effectively amplified: even if the initial comprehensive index is in a low range, as long as the high-frequency energy proportion reaches a certain level, the corrected comprehensive index will increase accordingly, thereby triggering the corresponding level of early warning in advance. This solves the problem of untimely identification of high-frequency related severe damage in the assessment. Simultaneously, the correction process is implemented through linear calculations, ensuring the simplicity of the assessment logic while allowing for adaptation to the sensitivity requirements of different equipment and components to high-frequency damage by adjusting the correction gain coefficient, further improving the specificity and reliability of the assessment method.

[0054] In one implementation, for the wind turbine main shaft bearing damage assessment scenario, the linear correction process for the comprehensive index under fault conditions is as follows: First, the value of the correction gain coefficient k was determined to be 0.8 through preliminary experimental calibration. This value was obtained by fitting the measured data of the wind turbine main shaft bearing at different damage stages, which ensures sensitive amplification of high-frequency crack signals and avoids misjudgment caused by overcorrection.

[0055] Secondly, calculate the overall high-frequency energy proportion. During equipment operation, when early microcracks appear in the bearing, crack propagation generates a large number of high-frequency acoustic emission signals. The ratio of the sum of the high-frequency band energy of all effective signals to the total energy of all frequency bands is calculated to obtain... The value reflects the dominance of high-frequency components in the current fault signal; the higher the proportion, the more significant the characteristics of high-risk crack-type damage.

[0056] The correction coefficient is calculated according to the linear correction logic: the correction gain coefficient k is compared with the overall high-frequency energy ratio. Multiply the results to obtain the high-frequency correction gain; add the calculated high-frequency correction gain to 1 to obtain the correction coefficient.

[0057] Finally, the corrected comprehensive index is obtained by multiplying the correction coefficient by the comprehensive index calculated in the above steps.

[0058] The revised composite index is compared with the composite index threshold. If the revised composite index is less than the first threshold... This indicates that the fault signal is weak and the key components have only minor abnormalities, such as slight surface friction or minor loosening, which have not had a substantial impact on the stability of equipment operation. There is no need to shut down the machine for treatment; it is only necessary to continuously track and monitor the trend of indicator changes.

[0059] If the corrected comprehensive index is greater than or equal to the first threshold And less than or equal to the second threshold If the damage is deemed to be progressive, the critical component is determined to be in a state of progressive damage. This state indicates that the component's damage has progressed beyond the initial minor abnormality stage and entered a continuous evolution process, with the fault characteristics showing a clear progressive trend. If this type of damage is not intervened in time, it will gradually evolve towards severe failure, which may lead to a decrease in equipment operational stability and an increase in the probability of downtime due to malfunctions.

[0060] If the revised comprehensive index is greater than the second threshold If the damage to the component has reached a critical stage, it indicates that there may be fatal defects such as through cracks or detachment of rolling elements. The machine must be stopped immediately for inspection and replacement to prevent the failure from escalating and causing equipment shutdown, production interruption, or even safety accidents.

[0061] This application also provides a damage assessment system for key components of low-speed heavy-load equipment, such as... Figure 2 As shown, it includes: The signal processing module is used to collect signals of acoustic emission events during the operation of key components of low-speed heavy-load equipment, filter noise from the signals to obtain time-domain signals, and perform frequency-domain transformation on the time-domain signals to obtain frequency-domain signals. The preliminary judgment module is used to divide the frequency domain signal into corresponding frequency bands according to the preset frequency band, calculate the energy proportion of each frequency band, and use the preset peak amplitude threshold and high frequency band energy proportion threshold as fault feature judgment conditions to judge the fault features. The feature parameter module is used to filter the time-domain and frequency-domain signals of acoustic emission events in fault states as valid signals. Based on each group of valid signals, the corresponding amplitude weight and frequency band weight are obtained. The weighted impact number, weighted energy and weighted count are calculated by combining the amplitude weight and frequency band weight of each group. The fault level determination module is used to substitute the weighted impact number, weighted energy, weighted count, and overall dominant frequency band type into the logarithmic linear determination model to calculate a comprehensive index. Based on the comprehensive index under healthy conditions, the fault level threshold is divided. The comprehensive index under fault conditions is linearly corrected by combining the overall high-frequency energy ratio, and the fault level is determined by the corrected comprehensive index.

[0062] This application also provides an electronic device, including: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by one or more processors, the one or more processors implement a damage assessment method for any critical component of a low-speed heavy-load equipment.

[0063] Specifically, the processor is an industrial-grade processor with strong real-time signal processing capabilities and multi-threaded computing performance. It can adapt to the rapid analysis requirements of transient fault signals in low-speed, heavy-load scenarios and efficiently complete core processes such as noise filtering, frequency domain transformation, weighted parameter calculation, and logarithmic linear judgment model calculation, ensuring that the processing delay of each set of acoustic emission event signals is controlled within an acceptable range for industrial applications.

[0064] The memory is used to store one or more programs and various key data. The programs include signal acquisition and control programs, noise filtering algorithm programs, frequency domain banding processing programs, characteristic parameter calculation programs, log-linear judgment model programs, and fault level judgment programs. The data includes preset benchmark parameters, health status sample data, real-time acquired raw signals, intermediate processing results, and historical evaluation records. The memory supports fast data reading and writing and long-term data retention, providing data support for dynamic threshold optimization and equipment operation and maintenance traceability.

[0065] When the program is executed by the processor, the processor will sequentially call the program of each module according to the logical steps of the damage assessment method, and collaboratively complete signal processing, feature extraction, model calculation and result output: First, the program controls the sensor interface to collect acoustic emission event signals of key components, calls the noise filtering algorithm to obtain the time domain signal and completes the frequency domain transformation; then, the frequency domain banding and energy ratio calculation program is executed, and the preliminary fault state is determined by combining the preset threshold; then, the effective signals are screened and the feature parameter calculation program is run to obtain the weighted impact number, weighted energy and weighted count; finally, the logarithmic linear judgment model program is called, and the comprehensive index is calculated by combining the overall dominant frequency band type. Based on the threshold of the health status data, after linear correction by the overall high frequency energy ratio, the accurate fault level judgment result is output, thus completely realizing the damage assessment method for key components of any low-speed heavy-load equipment.

[0066] The above embodiments are merely illustrative of the technical concept and features of the present invention, intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly, and should not be construed as limiting the scope of protection of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the present invention is defined by the appended claims rather than the foregoing description, and therefore all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention.

Claims

1. A damage assessment method for key components of low-speed heavy-load equipment, characterized in that, include: The system collects acoustic emission signals from key components of low-speed heavy-load equipment during operation, filters out noise to obtain time-domain signals, and performs frequency-domain transformation on the time-domain signals to obtain frequency-domain signals. The frequency domain signal is divided into corresponding frequency bands according to the preset frequency band, the energy proportion of each frequency band is calculated, and the preset peak amplitude threshold and high frequency band energy proportion threshold are used as fault feature judgment conditions to judge the fault features. The time-domain and frequency-domain signals of acoustic emission events in fault states are selected as valid signals. Based on each group of valid signals, the corresponding amplitude weight and frequency band weight are obtained. The weighted impact number, weighted energy, and weighted count are calculated by combining the amplitude weight and frequency band weight of each group. The weighted impact number, weighted energy, weighted count, and overall dominant frequency band type are substituted into the logarithmic linear judgment model to calculate the comprehensive index. The fault level threshold is divided based on the comprehensive index under the healthy state. The comprehensive index under the fault state is linearly corrected by combining the overall high-frequency energy ratio, and the fault level is judged by the corrected comprehensive index.

2. The damage assessment method for key components of low-speed heavy-load equipment according to claim 1, characterized in that, The method for obtaining the corresponding amplitude weight and frequency band weight based on each group of valid signals includes: The ratio of the peak amplitude of each valid signal group to the preset healthy amplitude is calculated as the amplitude weight of each group; the preset healthy amplitude is obtained based on the maximum allowable safe amplitude of the component. The frequency corresponding to the maximum spectral energy of each group of effective signals is taken as the main frequency, and the frequency band weight of each group is determined according to the frequency band where the main frequency is located.

3. The damage assessment method for key components of low-speed heavy-load equipment according to claim 2, characterized in that, The method for determining the band weight of each group based on the band where the main frequency is located is as follows: if the main frequency belongs to the dominant band of the effective signal in the group, the band weight is set to 1; if the main frequency does not belong to the dominant band, the band weight is a weight attenuation factor; the weight attenuation factor is a preset value; the dominant band of the effective signal in the group refers to the band with the largest energy proportion in the effective signal in the group.

4. The damage assessment method for key components of low-speed heavy-load equipment according to claim 1, characterized in that, The method for calculating the weighted impact number, weighted energy, and weighted count by combining the amplitude weight and frequency band weight of each group includes: There are n groups of valid signals. The weighted collision number is obtained by summing the products of the amplitude weight and the frequency band weight corresponding to the n groups of valid signals. The weighted energy is obtained by multiplying the energy of n groups of valid signals with their corresponding frequency band weights and summing the results. The frequency band weights corresponding to the n groups of valid signals are accumulated to obtain a weighted count.

5. The damage assessment method for key components of low-speed heavy-load equipment according to claim 1, characterized in that, The method for calculating the comprehensive index is as follows: the comprehensive index is obtained by summing the product of the overall dominant frequency band type code and the frequency band type correction coefficient, the product of the logarithm of the weighted impact number and the impact occurrence frequency weight coefficient, the product of the logarithm of the weighted energy and the damage energy weight coefficient, and the product of the logarithm of the weighted count and the activity level weight coefficient. The sum of the impact frequency weighting coefficient, the damage energy weighting coefficient, and the activity level weighting coefficient is 1.

6. The damage assessment method for key components of low-speed heavy-load equipment according to claim 5, characterized in that, The method for determining the overall dominant frequency band type coding is as follows: The frequency band energy percentage of all valid signals is statistically analyzed to obtain the overall low-frequency energy percentage, overall mid-frequency energy percentage, and overall high-frequency energy percentage. The frequency band corresponding to the largest frequency band energy percentage is taken as the overall dominant frequency band. When the overall dominant frequency band is a low frequency band, the corresponding overall dominant frequency band type code is 0; when the overall dominant frequency band is a mid frequency band, the corresponding overall dominant frequency band type code is 1; and when the overall dominant frequency band is a high frequency band, the corresponding overall dominant frequency band type code is 2.

7. The damage assessment method for key components of low-speed heavy-load equipment according to claim 1, characterized in that, The method for classifying fault level thresholds based on comprehensive indicators under health status is as follows: A comprehensive index set of acoustic emission events under all healthy conditions is calculated, and the mean and standard deviation of the comprehensive index set are taken. Two comprehensive index thresholds are set based on the mean and standard deviation. The comprehensive index thresholds are used to classify the fault levels, which include minor abnormalities, progressive damage, and severe damage.

8. The damage assessment method for key components of low-speed heavy-load equipment according to claim 1, characterized in that, The method for linearly correcting the comprehensive index under fault conditions by combining the overall high-frequency energy ratio is as follows: set a correction gain coefficient, use the correction gain coefficient to weight the overall high-frequency energy ratio, add the weighted overall high-frequency energy ratio to 1 to obtain the correction coefficient; The corrected comprehensive index is obtained by multiplying the correction factor by the comprehensive index.

9. A damage assessment system for key components of low-speed heavy-load equipment, characterized in that, The signal processing module is used to collect signals of acoustic emission events during the operation of key components of low-speed heavy-load equipment, filter noise from the signals to obtain time-domain signals, and perform frequency-domain transformation on the time-domain signals to obtain frequency-domain signals. The preliminary judgment module is used to divide the frequency domain signal into corresponding frequency bands according to the preset frequency band, calculate the energy proportion of each frequency band, and use the preset peak amplitude threshold and high frequency band energy proportion threshold as fault feature judgment conditions to judge the fault features. The feature parameter module is used to filter the time-domain and frequency-domain signals of acoustic emission events in fault states as valid signals. Based on each group of valid signals, the corresponding amplitude weight and frequency band weight are obtained. The weighted impact number, weighted energy and weighted count are calculated by combining the amplitude weight and frequency band weight of each group. The fault level judgment module is used to substitute the weighted impact number, weighted energy, weighted count, and overall dominant frequency band type into the logarithmic linear judgment model to calculate the comprehensive index. Based on the comprehensive index under the healthy state, the fault level threshold is divided. The comprehensive index under the fault state is linearly corrected by combining the overall high-frequency energy ratio, and the fault level is judged by the corrected comprehensive index.

10. An electronic device, characterized in that: The electronic device includes: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement a damage assessment method for key components of low-speed heavy-load equipment as described in any one of claims 1-8.

Images