Method for testing stress of small diesel engine driving gear based on contact pressure sensor

Abstract

The application discloses a small diesel engine transmission gear stress testing method based on contact pressure sensing, and particularly relates to the technical field of mechanical measurement and signal processing, and comprises the following steps: collecting original dynamic pressure signals reflecting gear meshing force through contact pressure sensors arranged on the bearing seat support surface of a gear box; filtering and denoising the signals, extracting effective gear meshing dynamic contact pressure signals, and calculating the characteristic values of the dynamic pressure signals of each meshing period; comparing the characteristic values with the reference pressure characteristic value spectrum under the pre-stored standard working condition in real time, identifying the current working condition of the engine, calling the pre-calibrated stress mapping model matched with the working condition, calculating the gear tooth root dynamic stress time series data, and through statistical cycle counting processing, generating the gear stress load spectrum used for fatigue life analysis. The application has the advantages of simple installation, strong anti-interference capability, working condition self-adaptation, fully automatic processing and the like, and improves the reliability and precision of stress testing.

Description

Technical Field

[0001] This invention relates to the field of mechanical measurement and signal processing technology, and more specifically, to a method for stress testing of small diesel engine transmission gears based on contact pressure sensing. Background Technology

[0002] Currently, the mainstream method for stress testing of transmission gears in small diesel engines is still based on resistance strain gauge technology. This method involves attaching strain gauges to the root of the gear teeth to sense the microscopic deformation of the gear under load, and then calculating the stress value based on the principles of materials mechanics. To achieve signal acquisition, slip rings or wireless telemetry devices are usually used to transmit the signals from the rotating parts to the static measurement system. In addition, some studies have used photoelasticity, acoustic emission, and other techniques for auxiliary analysis, or placed vibration sensors on the outside of the gearbox to indirectly assess the internal load state.

[0003] However, in practical use, it still has some drawbacks. For example, the installation process of resistance strain gauges is complex, and the requirements for gear structure space and surface treatment are extremely high. Moreover, the survival rate is low and the signal attenuation is severe under harsh working conditions such as high speed, oil contamination, and temperature changes. Slip rings suffer from contact wear and signal noise, while wireless transmission faces problems such as power supply difficulties, insufficient bandwidth, and electromagnetic interference, making it difficult to guarantee the reliability and accuracy of long-term measurements. It lacks the ability to dynamically adapt to complex and variable working conditions, and cannot identify the operating status and switch the corresponding calculation model in real time, resulting in a decrease in the accuracy of stress mapping. The degree of automation in data processing is low, relying on manual experience, and it is difficult to directly generate high-reliability load spectra. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for stress testing of small diesel engine transmission gears based on contact pressure sensing, which solves the problems mentioned in the background art through the following scheme.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for testing the stress of transmission gears in a small diesel engine based on contact pressure sensing, comprising:

[0006] S1: Acquire the original dynamic pressure signal reflecting the gear meshing force collected by the contact pressure sensor installed on the bearing support surface of the small diesel engine gearbox.

[0007] S2: The original dynamic pressure signal is filtered and denoised to extract the effective gear meshing dynamic contact pressure signal, and the dynamic pressure signal feature value for each meshing cycle is calculated from it. The dynamic pressure signal feature value includes peak value, mean value and pressure change slope.

[0008] S3: Compare the dynamic pressure signal feature value with the spectrum of the reference pressure feature value under different standard operating conditions stored in the database in real time, and identify and determine the current real-time operating condition of the engine based on the comparison result.

[0009] S4: Call the pre-calibrated stress mapping model that matches the real-time operating condition, substitute the pre-processed dynamic contact pressure signal into it, and calculate the dynamic stress time series data of the tooth root of the tested gear under the specific operating condition.

[0010] S5: Perform statistical cyclic counting on the dynamic stress time series data to generate a gear stress load spectrum for fatigue life analysis.

[0011] Preferably, the original dynamic pressure signal includes: a multi-channel voltage signal acquired by four piezoelectric pressure sensors symmetrically arranged on the bearing seats of the gearbox input shaft and output shaft, and the acquisition process uses a rotational speed pulse signal as an external trigger source to achieve synchronous acquisition.

[0012] Preferably, the gear meshing dynamic contact pressure signal includes: a reconstructed signal obtained by dual-frequency filtering of the original dynamic pressure signal through second-order Butterworth high-pass filtering and low-pass filtering, and denoising through 5-layer decomposition soft thresholding based on the sym8 wavelet basis.

[0013] Preferably, the dynamic pressure signal characteristic values ​​include: the peak value and the mean value obtained by calculating the signal for each meshing cycle, and the pressure change slope obtained by dividing the maximum value of the first-order difference of the signal by the sampling time interval.

[0014] Preferably, the real-time operating condition includes: calculating the Mahalanobis distance between the current feature value window statistics and the pre-stored baseline operating condition spectrum, and finding the operating condition corresponding to the minimum distance value being lower than a set threshold.

[0015] Preferably, the pre-calibrated stress mapping model includes: a model independently calibrated based on different working conditions, which may be in the form of a neural network model, a transfer function model or a nonlinear regression model, and includes signal phase compensation, temperature parameter adjustment and load scaling mechanism during calculation.

[0016] Preferably, the dynamic stress time series data includes: stress time history data, synchronous time series, operating condition identifiers, and quality assessment indicators based on model confidence and input signal quality.

[0017] Preferably, the statistical cycle counting process includes: using a three-parameter rainflow counting method to identify and extract stress cycles, calculating the stress amplitude, average stress, and cycle number of each cycle, and performing small load elimination and average stress correction on the generated load spectrum.

[0018] Preferably, the gear stress load spectrum includes: matrix structure data arranged from largest to smallest stress amplitude, each row containing three main data columns: stress amplitude, average stress, and number of cycles, and is accompanied by metadata such as total number of cycles, maximum stress amplitude, and working condition identifier.

[0019] The technical effects and advantages of this invention are as follows:

[0020] Excellent installation compatibility and low implementation threshold: No need to modify the gear body, the sensor is placed on the support surface of the gearbox bearing seat, with clear installation parameters and calibration procedures, simplifying the installation process, adapting to the compact space of small diesel engines, and avoiding damage to the integrity of the transmission system.

[0021] Strong anti-interference capability and more reliable measurement: Through the dual guarantee of low-noise hardware transmission and filtering and wavelet denoising software processing, combined with real-time signal quality monitoring, the problems of low strain gauge survival rate and high transmission noise under harsh working conditions are solved, ensuring long-term measurement stability;

[0022] Adaptive to working conditions and high mapping accuracy: Based on the real-time identification of working conditions using Mahalanobis distance, the system automatically calls up the matching stress mapping model and combines it with a triple compensation mechanism of phase, temperature and load to break through the limitations of traditional fixed models and improve the accuracy of stress calculation under varying working conditions.

[0023] The process is fully automated and the results are highly practical: it realizes closed-loop automation from signal acquisition and preprocessing to load spectrum generation, without the need for manual intervention, and directly outputs the corrected structured stress load spectrum, which can accurately support fatigue life analysis.

[0024] Good robustness and scalability: Wired transmission avoids wear and tear risks, and long-term accuracy is guaranteed through database self-learning and model recalibration mechanisms. The model library can be flexibly expanded to adapt to different testing needs and equipment models. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall structure of the present invention.

[0026] Figure 2 This is a schematic diagram of the signal processing and feature extraction structure of the present invention.

[0027] Figure 3 This is a schematic diagram of the working condition identification and stress mapping structure of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] refer to Figures 1-3 The stress testing method for small diesel engine transmission gears based on contact pressure sensing, as shown, includes:

[0030] S1: Acquire the original dynamic pressure signal reflecting the gear meshing force collected by the contact pressure sensor installed on the bearing support surface of the small diesel engine gearbox.

[0031] S2: The original dynamic pressure signal is filtered and denoised to extract the effective gear meshing dynamic contact pressure signal, and the dynamic pressure signal feature value for each meshing cycle is calculated from it. The dynamic pressure signal feature value includes peak value, mean value and pressure change slope.

[0032] S3: Compare the dynamic pressure signal feature value with the spectrum of the reference pressure feature value under different standard operating conditions stored in the database in real time, and identify and determine the current real-time operating condition of the engine based on the comparison result.

[0033] S4: Call the pre-calibrated stress mapping model that matches the real-time operating condition, substitute the pre-processed dynamic contact pressure signal into it, and calculate the dynamic stress time series data of the tooth root of the tested gear under the specific operating condition.

[0034] S5: Perform statistical cyclic counting on the dynamic stress time series data to generate a gear stress load spectrum for fatigue life analysis.

[0035] S1: Through a properly installed sensor system, a high-quality, distortion-free original voltage signal that can accurately reflect the magnitude of the gear meshing force is acquired.

[0036] Two piezoelectric pressure sensors are symmetrically installed on the bearing housing support surfaces of the input and output shafts of the gearbox, for a total of four measuring points. Before installation, the mounting surfaces are thoroughly cleaned with an organic solvent to ensure surface roughness. Less than or equal to 1.6 ; Use a special installation fixture with 23 By 27 The torque between them fixes the sensor, ensuring that the sensor is in full contact with the support surface and that the force is evenly distributed;

[0037] Immediately after installation, on-site calibration should be performed. Static calibration should be conducted using a standard pressure source at 20%, 40%, 60%, 80%, and 100% of the measurement range, and the sensitivity coefficient at each measurement point should be recorded. Subsequently, dynamic frequency response testing should be performed using the impact hammer method to ensure that the frequency response consistency error of each channel is less than 3%.

[0038] The sensor is connected to a charge amplifier via a low-noise shielded cable, with the amplifier gain set to 4mV / N and the low-pass filter cutoff frequency set to 10kHz. The signal is then input to a 16-bit high-precision data acquisition card with a sampling rate set to 25.6kHz to ensure complete capture of the high-frequency dynamic signals of gear meshing.

[0039] After the diesel engine reaches a stable operating state, data collection begins, including: idling condition (800 rpm, 0% load); medium load condition (1500 rpm, 50% load); rated condition (2200 rpm, 100% load); continuous data collection under each condition includes at least 30 complete gear meshing cycles.

[0040] The rotational speed pulse signal is used as an external trigger source to ensure that the acquisition is synchronized with the shaft rotation. During the acquisition process, the signal-to-noise ratio, peak amplitude and waveform integrity of each channel signal are monitored in real time. The signal-to-noise ratio should be greater than 40dB and the peak amplitude should be between 20% and 80% of the range. If an abnormal signal is found, the acquisition should be interrupted immediately and the fault should be eliminated.

[0041] The acquired multi-channel raw voltage signals are first subjected to mean zeroing processing to eliminate the influence of DC components and zero-point drift; the processed signal is the raw dynamic pressure signal.

[0042]

[0043] in Indicates the first Time-domain voltage signals acquired by each sensor channel; It is a time variable;

[0044] The raw dynamic pressure signal is output in real time as a binary data stream, and the following metadata is embedded simultaneously:

[0045] Sensor parameters: sensitivity of each sensor (e.g., 4mV / N) and specific installation location markings;

[0046] Acquisition system parameters: sampling rate (25.6kHz), range and gain settings for each channel;

[0047] Operating parameters: Real-time monitoring data such as speed, load torque, and oil temperature;

[0048] Timestamp: Collects time-domain information such as start time and duration;

[0049] It should be further noted that the sensor mounting torque range is 23. By 27 The selection of torque is based on the optimal range determined through extensive experimental verification. Too low a torque will result in insufficient contact between the sensor and the support surface, introducing measurement errors; too high a torque may cause structural deformation of the support surface or the sensor itself, affecting signal accuracy.

[0050] Surface roughness Less than or equal to 1.6 It achieves the best balance between cost, processing difficulty and measurement performance, ensuring no gaps or impurities between the sensor and the support surface, and guaranteeing the linearity and consistency of pressure transmission;

[0051] The sampling rate was set to 25.6kHz to satisfy the Nyquist sampling theorem, ensuring that the highest frequency components that may occur during gear meshing can be fully captured and avoiding aliasing.

[0052] S2: The output of the original dynamic pressure signal sequence The process involves extracting valid gear meshing signals and calculating key feature values.

[0053] First of all Preprocessing is performed: a second-order Butterworth high-pass filter with a cutoff frequency of 10kHz is used to eliminate low-frequency vibration interference, and a second-order Butterworth low-pass filter with a cutoff frequency of 100Hz is used to remove high-frequency noise; the filtered signal is denoted as the preprocessed signal. ;

[0054] Preprocessed signal Wavelet denoising processing is performed: a 5-level decomposition is conducted using the sym8 wavelet basis. The threshold for each level is determined using the Stein unbiased risk estimation criterion. After soft thresholding of the detail coefficients, the signal is reconstructed. This process generates an effective gear meshing dynamic contact pressure signal, denoted as: This signal has effectively removed environmental noise and electromagnetic interference, while preserving the complete dynamic characteristics of gear meshing.

[0055] Based on the rotational speed pulse signal, The gear is divided according to its meshing cycle; each complete meshing cycle corresponds to one rotation of the gear, and the resulting division yields... One periodic signal segment: ;

[0056] Each of them Representing the Dynamic pressure signal for each meshing cycle;

[0057] For each period signal Calculate the following dynamic pressure signal characteristic values:

[0058] Peak value: ; Calculate the maximum amplitude of the signal within this period, reflecting the maximum meshing impact force.

[0059] Mean: ;in This represents the number of sampling points within the cycle, reflecting the average meshing pressure during that cycle.

[0060] Pressure change slope:

[0061] The maximum value of the first-order difference of the signal is calculated and divided by the sampling time interval. The maximum rate of pressure change is obtained, reflecting the dynamic characteristics of the meshing process.

[0062] Ultimately, two levels of products are obtained:

[0063] Level 1: The set of dynamic pressure signal characteristic values ​​for each meshing cycle, denoted as:

[0064]

[0065] Each of them Including the Three characteristic values ​​for each period.

[0066] Level 2: The eigenvalue sequences of all periods form three characteristic time series:

[0067] Peak sequence:

[0068] Mean sequence:

[0069] Slope sequence:

[0070] The three feature sequences fully characterize the dynamic properties of the gear meshing process, providing accurate input features for subsequent working condition identification and stress mapping.

[0071] It should be further explained that the symmetry and tight support characteristics of the sym8 wavelet gene used can effectively match the waveform of the gear meshing impact signal, reducing reconstruction distortion. The setting of 5-level decomposition is the best balance between computational efficiency and signal detail preservation, ensuring that the main noise components can be separated while retaining the key information of the meshing frequency and its harmonics. The Stein unbiased risk estimation criterion is an adaptive threshold selection method. Its core principle is to minimize the prediction mean square error risk between the estimated signal and the real signal. Compared with general thresholding methods, this criterion can dynamically adjust the threshold according to the actual noise level of each layer of the signal, which is especially suitable for non-stationary gear dynamic signals and has a better denoising effect.

[0072] S3: Based on dynamic pressure signal characteristic values By comparing the current operating conditions of the engine with the pre-stored reference spectrum in real time, the engine's current operating conditions can be identified.

[0073] Pre-established includes A baseline database of standard operating conditions, for each operating condition A spectrum corresponding to a reference pressure characteristic value :

[0074]

[0075] in and These represent the mean and standard deviation of each characteristic value under this operating condition, obtained through statistical analysis of a large amount of experimental data. Standard operating conditions include, but are not limited to, typical operating states such as idling, medium load, rated load, and overload.

[0076] Using sliding window technology, the latest A window of characteristic values ​​for each meshing cycle:

[0077] Perform statistical analysis and calculate the statistics of the feature values ​​within the window:

[0078]

[0079] Calculate the current window statistic using Mahalanobis distance. With each benchmark operating condition Similarity between them:

[0080]

[0081] in For the baseline operating condition The covariance matrix.

[0082] Find the benchmark condition most similar to the current window: ,in For variables The independent variable that minimizes the value of the variable;

[0083] If the minimum Mahalanobis distance If the value is less than a set threshold (e.g., 3.0), then the current operating condition is determined to be... Otherwise, mark it as an unknown operating condition or a transitional operating condition.

[0084] The final output, generated through decision-making logic, is the current operating condition identifier.

[0085] in It is structured data containing the following fields:

[0086] Operating condition type: enumerated values, such as idling, medium load, rated load, overload, unknown;

[0087] Confidence level: Match confidence level (0-100%) calculated based on Mahalanobis distance;

[0088] Timestamp: Time of condition determination;

[0089] Feature matching degree: The individual matching degree of each feature dimension.

[0090] The system also includes a feedback mechanism: when the same operating condition remains stable for a period of time (e.g., 100 consecutive cycles) and the confidence level is greater than 95%, the current window feature statistics are used to update the corresponding baseline operating condition spectrum. This enables online optimization and learning of the database.

[0091] It should be further noted that the square of the Mahalanobis distance follows... Distribution; threshold 3.0 corresponds to The distribution has a high confidence level (approximately 99.7%), meaning that if the current data does indeed belong to a certain operating condition, its Mahalanobis distance is likely to be less than this value. Exceeding this threshold provides statistical evidence to reject the null hypothesis of a difference in operating conditions. This threshold effectively balances the risks of misjudgment and missed judgment: a threshold of 2.0 easily misjudges normal fluctuations as changes in operating conditions, leading to frequent system switching; a threshold of 4.0 is insensitive to genuine anomalies. A value of 3.0 is widely applicable to various pattern recognition and anomaly detection scenarios, combining robustness to normal fluctuations with sensitivity to abnormal conditions, ensuring the accuracy and reliability of subsequent stress mapping.

[0092] It should be further noted that a confidence level greater than 95% indicates that the probability of misjudgment in identifying the current operating condition is less than 5%. This high threshold ensures extremely high reliability of the input data, effectively filters out transient interference signals, prevents erroneous data from polluting the benchmark database, and guarantees the accuracy of the learning mechanism and the long-term stability of the core model.

[0093] S4: Based on effective gear meshing dynamic contact pressure signal and real-time operating status indicators By calling the corresponding stress mapping model, the dynamic stress sequence data of the gear tooth root under this specific working condition can be calculated.

[0094] Establish a pre-calibrated stress mapping model library containing multiple working conditions, for each working condition (Such as idling, medium load, rated load, overload, etc.) corresponds to an independently calibrated stress mapping model. The model is represented as follows:

[0095]

[0096] in, For working conditions The corresponding stress mapping model;

[0097] For working conditions The calibrated mapping function can be:

[0098] Neural network models (such as BP neural networks and RBF neural networks) have the advantages of strong nonlinear fitting ability and adaptability to complex mapping relationships;

[0099] Transfer function model: It is applicable to linear or approximately linear systems;

[0100] Nonlinear regression model: It can capture the dynamic changes in stress; among them, The coefficients of the linear terms; The coefficient of the quadratic term; The coefficients of the derivative term;

[0101] According to real-time operating status indicators The load condition type field automatically retrieves the corresponding stress mapping model from the model library.

[0102]

[0103] in The operating condition number is the one corresponding to the current operating condition type. If the current operating condition type is unknown, the default conservative model (corresponding to the worst operating condition) is used.

[0104] The pre-processed effective gear meshing dynamic contact pressure signal Input the selected stress mapping model to calculate the dynamic stress at the tooth root in real time:

[0105] During the calculation process, multiple compensation mechanisms are introduced to improve accuracy:

[0106] Signal phase compensation: Based on the transmission path between the sensor installation position and the tooth root, the signal phase is compensated. A fixed phase offset is applied to correct for signal delay and ensure the accuracy of the stress peak time;

[0107] Temperature compensation: The model parameters are dynamically adjusted based on the real-time monitored lubricating oil temperature to reflect the change of the material's elastic modulus with temperature.

[0108] Load compensation: The stress amplitude is scaled proportionally based on the real-time torque signal to ensure the accuracy of calculations under different load conditions.

[0109] The time series of dynamic stress at the gear tooth root is finally obtained through the calculation process: The dynamic stress time series is a structured data set that includes:

[0110] Stress time history data: ;

[0111] Time series: , and input signal Keep in sync

[0112] Operating condition information: Inherited from Operating condition indicators and confidence level

[0113] Quality Indicators: Quality assessment metrics based on model confidence and input signal quality

[0114] System real-time monitoring model output Compared with actual measured value The deviation is considered, and when the deviation continues to exceed a threshold, the model update procedure is initiated.

[0115] if If the value exceeds the threshold, model recalibration is triggered.

[0116] It should be further noted that the model recalibration trigger threshold is set to... The relative error, i.e. At this time, the model update procedure is initiated. This threshold is higher than the inherent comprehensive error of high-precision stress measurement systems (typically 100%). This effectively avoids false triggering caused by measurement noise and accidental interference, ensuring the necessity and seriousness of the recalibration action; The deviation clearly exceeds the expected accuracy range of the stress mapping model, indicating that the model has experienced substantial drift or failure, and continued use will lead to unacceptable errors in the fatigue life prediction results.

[0117] S5: Based on the time series data of dynamic stress at the gear tooth root Through statistical cyclic counting, a gear stress load spectrum for fatigue life analysis is generated.

[0118] For the input dynamic stress time series Data preprocessing was performed. A moving average filtering algorithm was used to smooth the raw stress data, with a window width set to 5-10 data points to eliminate high-frequency noise interference while preserving important stress fluctuation characteristics. Subsequently, a peak detection algorithm based on slope change was used to identify all extreme points (including maxima and minima) in the stress time series. The peak detection algorithm accurately extracts stress peaks and troughs by calculating the first-order difference of continuous data points and identifying points where the sign of the difference value changes. The extracted sequence of extreme points is denoted as:

[0119] ,in This represents the total number of extreme points, each with a timestamp.

[0120] The three-parameter rainflow counting method was used to analyze the extreme value sequence. Cyclic counting is performed; the rainflow counting method is based on the stress-strain hysteresis loop principle of materials, which can accurately identify and extract complex stress cycles; the counting process first rearranges the extreme value sequence into a positive sequence starting from the maximum value, and then extracts the complete stress cycles in sequence according to the rules of the rainflow method. For each identified stress cycle, three key parameters are calculated:

[0121] stress amplitude , indicating the intensity of the cyclic fluctuation;

[0122] Mean stress , indicating the static stress level of the cycle;

[0123] The initial number of loop iterations is set to 1.

[0124] Special attention is paid to cycle identification under variable amplitude load conditions during the processing to ensure that small cycles nested within large cycles can be accurately extracted.

[0125] All stress cycles identified by the rainflow counting method are grouped and statistically analyzed according to their stress amplitude. First, the entire stress amplitude range is determined, and the difference between the maximum and minimum stress amplitudes is calculated. An equally spaced grouping method is then used to divide the stress amplitude range into... A number of intervals, usually the number of intervals Based on the stress range and data accuracy requirements, the number of groups is set to 20-50, with the width of each interval being [missing information]. ;

[0126] The total number of stress cycles within each stress interval is calculated, and the average stress level of the cycles within that interval is also recorded. For each stress interval, the statistical distribution characteristics of the number of cycles are also calculated, including the mean, variance, and confidence interval.

[0127] Statistical processing was used to generate the final gear stress-load spectrum. The load spectrum is represented by a matrix structure, with each row representing a stress level range, containing three main data columns: stress amplitude. Mean stress and number of loops The load spectrum is arranged in descending order of stress amplitude. The load spectrum product also includes metadata information, such as total number of cycles, maximum stress amplitude, statistical characteristics of average stress level, data acquisition time period, and operating condition identifier.

[0128] The initially generated gear stress load spectrum undergoes a small load rejection process. Based on the fatigue limit characteristics of the material, small load cycles with stress amplitudes below 30% of the fatigue limit are removed from the load spectrum. Then, mean stress correction is performed. Using mean stress correction criteria such as Goodman or Gerber, non-zero mean stress cycles are transformed into equivalent zero mean stress cycles. Finally, the load spectrum is accumulated. When analyzing data from multiple consecutive time periods, the load spectra of the corresponding time periods are accumulated according to stress levels to form a cumulative load spectrum representing the entire operating cycle.

[0129] It should be further clarified that the three parameters in the aforementioned three-parameter rainflow counting method specifically refer to the amplitude of the stress cycle ( ), mean ( ) and number of cycles ( The core advantage of this method lies in its ability to accurately identify and extract closed stress-strain hysteresis loops that contribute to material fatigue damage from complex random load time histories, and its counting results are more consistent with the physical mechanism of material fatigue damage.

[0130] It should be further explained that the number of groups into which the stress amplitude range is divided... Setting the number of groups to 20-50 is a typical choice based on engineering practice; too few groups (e.g., too many groups) will result in insufficient load spectrum resolution, masking certain damaging stress levels; too many groups (e.g., too few groups) will result in insufficient load spectrum resolution, masking certain damaging stress levels; This would result in an overly sparse statistical distribution, with too few iterations in each interval, reducing the stability of subsequent fatigue damage calculations.

[0131] It should be further explained that setting the small load rejection threshold to 30% of the material's fatigue limit is an engineering simplification based on Miner's linear cumulative damage theory. Load cycles with stress levels below a certain percentage (such as 30%) of the fatigue limit have negligible contribution to the total damage. This can compress the amount of load spectrum data, improve the efficiency of subsequent fatigue analysis, and the impact on the accuracy of the final life prediction results is within an acceptable engineering error range.

[0132] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0133] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for testing the stress of a small diesel engine drive gear based on contact pressure sensing, characterized by, The method comprises the following steps: S1: acquiring original dynamic pressure signals reflecting gear engagement force collected by contact pressure sensors arranged on the bearing seat support surface of a small diesel engine gear box; S2: filtering and denoising the original dynamic pressure signals to extract valid gear engagement dynamic contact pressure signals, and calculating dynamic pressure signal characteristic values of each engagement cycle, including peak value, mean value and pressure change slope; S3: comparing the dynamic pressure signal characteristic values with reference pressure characteristic value spectrum of different standard working conditions pre-stored in a database in real time, and identifying and determining the current real-time running working condition of the engine according to the comparison result; S4: calling a pre-calibrated stress mapping model matched with the real-time running working condition, substituting the pre-processed dynamic contact pressure signals, and calculating dynamic stress time series data of the measured gear tooth root under the real-time running working condition; S5: performing statistical cycle counting processing on the dynamic stress time series data to generate a gear stress load spectrum for fatigue life analysis.

2. The compact diesel engine transmission gear stress testing method based on contact pressure sensing according to claim 1, characterized by, The original dynamic pressure signals include multi-channel voltage signals collected by four piezoelectric pressure sensors symmetrically arranged on the bearing seat support surface of the input shaft and output shaft of the gear box, and the collection process uses a rotating speed pulse signal as an external trigger source to realize synchronous collection.

3. The compact diesel engine drive gear stress testing method based on contact pressure sensing according to claim 1, characterized by, The gear engagement dynamic contact pressure signals include reconstructed signals obtained after the original dynamic pressure signals are subjected to double-frequency filtering of second-order Butterworth high-pass filtering and low-pass filtering, and 5-layer decomposition soft threshold denoising processing based on a sym8 wavelet basis.

4. The compact diesel engine transmission gear stress testing method based on contact pressure sensing according to claim 1, characterized by, The dynamic pressure signal characteristic values include peak value and mean value obtained by calculating each engagement cycle signal, and a pressure change slope obtained by dividing the maximum value of the first-order difference of the signal by the sampling time interval.

5. The compact diesel power transmission gear stress testing method based on contact pressure sensing according to claim 1, characterized in that, The real-time running working condition is determined by calculating the Mahalanobis distance between the current characteristic value window statistics and the pre-stored reference working condition spectrum, and finding the working condition corresponding to the minimum distance value below the set threshold.

6. The compact diesel power transmission gear stress testing method based on contact pressure sensing according to claim 1, characterized in that, The pre-calibrated stress mapping model includes a model calibrated independently based on different working conditions, which includes a neural network model, a transfer function model or a nonlinear regression model, and the calculation contains signal phase compensation, temperature parameter adjustment and load scaling mechanism.

7. The compact diesel power transmission gear stress testing method based on contact pressure sensing according to claim 1, characterized in that, The dynamic stress time series data includes stress time course data, synchronous time series, working condition identification and quality evaluation flag based on model confidence and input signal quality.

8. The compact diesel power transmission gear stress testing method based on contact pressure sensing of claim 1, wherein, The statistical cycle counting processing includes using a three-parameter rainflow counting method to identify and extract stress cycles, calculating the stress amplitude, average stress and cycle number of each cycle, and performing small load elimination and average stress correction on the generated load spectrum.

9. The compact diesel power transmission gear stress testing method based on contact pressure sensing of claim 1, wherein, The gear stress load spectrum includes matrix structure data arranged in descending order of stress amplitude, each row containing three main data columns of stress amplitude, average stress and cycle number, and total cycle number, maximum stress amplitude and working condition identification metadata.

Images