A heavy haul locomotive life prediction method based on regression equation and related system

Abstract

The application discloses a heavy haul locomotive life prediction method based on a regression equation, collects measured alternating stress time series data of a bearing component of a heavy haul locomotive; takes a probability P-S-N fatigue curve as a single cycle fatigue life calculation benchmark, takes the collected measured alternating stress time series data as input, and constructs a modified Corten-Dolan nonlinear cumulative damage model with a load interaction correction term; calculates fatigue damage in different service stages based on the nonlinear cumulative damage model, and fits to obtain a service parameter-damage regression equation; and calculates the remaining fatigue life of the component based on the service parameter-damage regression equation. The application has the characteristics of realizing accurate prediction of the life of the heavy haul locomotive frame, high life prediction accuracy, and accurate quantification of nonlinear damage evolution.

Description

Technical Field

[0001] This invention relates to the field of heavy-duty locomotive technology, and in particular to a method and system for predicting the lifespan of heavy-duty locomotives based on regression equations. Background Technology

[0002] As a critical component of trains, bogies play a vital role in guiding, traction, braking, and load-bearing, significantly impacting passenger comfort and reliability. Numerous studies have shown that bogie failure is primarily caused by fatigue, affecting vehicles including, but not limited to, subways, freight locomotives, and high-speed trains. With the expansion of global railway networks and the increase in train speeds, the service environment faced by bogies is becoming increasingly complex and demanding, making accurate assessment of bogie fatigue strength particularly crucial.

[0003] In related technologies, the fatigue reliability of bogie frames is often assessed using three methods: bogie bench tests, track tracking tests, and finite element simulations. However, since bench tests and finite element simulations are theoretical calculations, they lack practical feasibility. Therefore, track test data is often used to verify the reliability of the bogie frame. For example, related technologies use quasi-static theory to identify and separate dynamic stress signals from high-speed train bogie frames, thereby establishing a load spectrum with consistent damage. Furthermore, kernel density estimation and quasi-static measurement methods based on sample characteristics are used to establish a standardized load spectrum, and the accuracy of this method in assessing structural fatigue reliability is verified through measured track loads.

[0004] However, regarding the fatigue life assessment methods for heavy-duty locomotive bogie frames in related technologies, on the one hand, bench tests and simulation results are detached from actual service conditions and cannot reflect the true dynamic stress characteristics of track operation. Furthermore, the assessment results of the traditional Miner linear cumulative damage model are too conservative and cannot accurately quantify nonlinear damage evolution. On the other hand, the standard load spectrum method does not consider the discreteness of material parameters and the differences in measured data, resulting in large life prediction errors. Moreover, existing methods have not established a direct correlation model between operating mileage and fatigue damage, making it difficult to achieve dynamic prediction of remaining life based on service data. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for predicting the life of heavy-duty locomotives based on regression equations, which overcomes the shortcomings of existing methods for predicting the life of heavy-duty locomotives, such as large life prediction errors and difficulty in dynamically predicting the remaining life based on service data.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for predicting the lifespan of heavy-haul locomotives based on regression equations, comprising the following steps: S1. Collect measured alternating stress time series data of the load-bearing components of heavy-duty locomotives; S2. Using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation, and taking the measured alternating stress time series data collected in step S1 as input, a modified Corten-Dolan nonlinear cumulative damage model with load interaction correction terms is constructed. The expression of the nonlinear cumulative damage model is as follows: ;in, This indicates fatigue damage to the structure. This represents the actual number of cycles corresponding to each stress order. This represents the fatigue life corresponding to each stress level. Representing the Level stress amplitude, and These represent the maximum stress and its corresponding fatigue life among a series of loads applied to the specimen. The stress level; These are material parameters; the load interaction correction term is used to characterize the accelerated deterioration effect of high-amplitude alternating loads on component fatigue; S3. Based on the nonlinear cumulative damage model, calculate the fatigue damage at different service stages and fit the service parameter-damage regression equation. S4. Calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

[0007] The service parameters mentioned above refer to quantitative parameters characterizing the fatigue accumulation process of components, including at least one of running mileage, running time, and cumulative load. The present invention preferably uses running mileage as the service parameter, and a running mileage-damage regression equation is formed accordingly.

[0008] This invention uses measured alternating stress data of components as the basis for calculation, takes into account the performance dispersion of batch components through probabilistic PSN fatigue benchmark, restores the fatigue deterioration mechanism with nonlinear damage model with load interaction correction, and finally constructs a life prediction channel with service parameters-damage regression equation. It significantly improves the accuracy of life prediction while greatly reducing the amount of online calculation, and adapts to the engineering needs of heavy-duty locomotives for dynamic prediction of remaining life.

[0009] The modified Corten-Dolan model is used as the nonlinear damage calculation carrier. The load interaction correction term can be adapted to different component materials and working conditions through experimental calibration. The model has strong versatility, and the damage calculation accuracy is more in line with the actual load action law of heavy-duty locomotives.

[0010] Furthermore, the specific implementation process of step S3 includes: dividing the entire service range of the component, segmenting the alternating stress time series data according to the service range, inputting the alternating stress data of each segment into the nonlinear cumulative damage model, calculating the cumulative fatigue damage corresponding to each segment; and fitting the service parameter-damage regression equation with the service parameter as the independent variable and the segmented cumulative fatigue damage as the dependent variable.

[0011] By adopting a segmented approach based on service intervals, the load distribution characteristics of different service stages are matched to avoid biases caused by uneven load distribution in global fitting, thereby further improving the fitting accuracy of the regression equation and the reliability of life prediction.

[0012] Furthermore, the expression for the service parameter-damage regression equation is: ;in, Indicates service parameters (preferably operating mileage). , , The regression coefficient is determined by fitting the measured damage-service parameter data using an orthogonal distance regression algorithm.

[0013] Furthermore, the expression for the remaining fatigue life is: ;in, The total service life mileage is obtained by inversely solving the service parameter-damage regression equation by setting the value of the service parameter to the fatigue threshold: (Take the positive root with a reasonable physical meaning). The fatigue threshold, This indicates the locomotive's current cumulative mileage.

[0014] Furthermore, the method for predicting the lifespan of heavy-haul locomotives also includes: generating maintenance guidance information for heavy-haul locomotives based on the remaining fatigue life of components calculated in step S4, wherein the maintenance guidance information includes at least one of component maintenance priority and locomotive speed limit control instructions.

[0015] By transforming lifespan prediction results into executable operation and maintenance control instructions, a complete industrial closed loop of "data acquisition - calculation and analysis - operation and maintenance control" is formed, enhancing the engineering application value of this invention and directly serving the operation and maintenance management of heavy-duty locomotives.

[0016] As an inventive concept, this invention also provides a heavy-duty locomotive life prediction system based on regression equations, comprising: The data acquisition module is used to collect measured alternating stress time-series data of the load-bearing components of heavy-duty locomotives; The damage model construction module is used to construct a modified Corten-Dolan nonlinear cumulative damage model with load interaction correction terms, using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation and the measured alternating stress time series data collected by the data acquisition module as input. The regression fitting module is used to calculate fatigue damage at different service stages based on the nonlinear cumulative damage model and fit the service parameter-damage regression equation. The life prediction module is used to calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

[0017] As an inventive concept, the present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the steps of the above method.

[0018] As an inventive concept, the present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon; when the computer program / instructions are executed by a processor, they implement the steps of the above-described method.

[0019] As an inventive concept, the present invention also provides a computer program product, including a computer program / instructions; when the computer program / instructions are executed by a processor, they implement the steps of the above-described method.

[0020] Compared with the prior art, the beneficial effects of this invention are as follows: by using measured dynamic stress data of the line, using the PSN curve as the fatigue benchmark, and combining the modified Corten-Dolan nonlinear cumulative damage model, the cumulative damage under variable amplitude load is accurately calculated, and then the service parameter-damage regression equation is established to achieve accurate prediction of the life of the heavy-haul locomotive frame. The life prediction accuracy is high, and the nonlinear damage evolution can be accurately quantified. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the dynamic stress acquisition system provided in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the distribution of measuring points on the bogie frame provided in Embodiment 1 of the present invention; Figure 4 The waveform and time-frequency diagram of the dynamic stress signal at measuring points (1-12) of the vibration damper installed on the crossbeam provided in Embodiment 1 of the present invention; Figure 5 The waveform and time-frequency diagram of the dynamic stress signal at measuring points (1-24) of the motor hanger provided in Embodiment 1 of the present invention; Figure 6The waveform and time-frequency diagram of the dynamic stress signal at measuring points (1-25) of the gearbox hanger provided in Embodiment 1 of the present invention; Figure 7 The dynamic stress signal waveform and time-frequency diagram of the brake hanger measuring points (1-29) provided in Embodiment 1 of the present invention; Figure 8 The waveform and time-frequency diagram of the dynamic stress signal at measuring points (1-46) at the connection between the traction rod and the crossbeam provided in Embodiment 1 of the present invention; Figure 9 The waveform and time-frequency diagram of the dynamic stress signal at the positioning arm measuring point (1-51) provided in Embodiment 1 of the present invention; Figure 10 The waveform and time-frequency diagram of the dynamic stress signal at the positioning arm measuring point (1-55) provided in Embodiment 1 of the present invention; Figure 11 The dynamic stress signal waveform and time-frequency diagram of the gearbox hanger measuring points (1-56) provided in Embodiment 1 of the present invention; Figure 12 The waveform and time-frequency diagram of the dynamic stress signal at the measuring points (1-65) of the anti-hunting vibration damper provided in Embodiment 1 of the present invention; Figure 13 This is a scatter plot of the running mileage and damage at each measuring point provided in Embodiment 1 of the present invention; Figure 14 This is a radar image showing the damage comparison at each measuring point provided in Embodiment 1 of the present invention; Figure 15 This is a comparison chart of the predicted lifetimes of various measuring points provided in Embodiment 1 of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0023] Example 1 This embodiment provides a method for predicting the lifespan of heavy-haul locomotives based on regression equations, such as... Figure 1 As shown, the steps include: S100: Collect measured alternating stress time series data of load-bearing components of heavy-duty locomotives.

[0024] S200: Using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation, a nonlinear cumulative damage model with load interaction correction terms is constructed.

[0025] S300: Based on the aforementioned nonlinear cumulative damage model, calculate the fatigue damage at different service stages and fit the service parameter-damage regression equation.

[0026] S400: Calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

[0027] The measured alternating stress time series data were obtained by performing dynamic stress line tracking tests on key parts of the bogie frame of heavy-duty locomotives using strain gauge method.

[0028] The dynamic stress acquisition system consists of the following components: Figure 2 As shown, the circuit testing equipment includes resistance strain gauges, data transmission connection lines, an eDAQ multi-channel data acquisition instrument, and a GPS acquisition module. The resistance value of the strain gauge is 119.8±0.1Ω, and the sensitivity coefficient is 2.17±1%. It adopts the Wheatstone full-bridge circuit method to improve sensitivity and eliminate the effect of temperature difference.

[0029] like Figure 3 As shown, in the dynamic stress line tracking test, nine key stress areas on the bogie frame were selected for dynamic stress acquisition tests. The specific locations and corresponding numbers of each measuring point are as follows: anti-hunting damper seat measuring point (1-65), damper mounted on the crossbeam measuring point (1-12), motor hanger measuring point (1-24), gearbox hanger measuring point (1-56, 1-25), positioning swing arm measuring point (1-51, 1-55), brake hanger measuring point (1-29), and traction rod and crossbeam connection measuring point (1-46).

[0030] The test section for this line tracking test was selected as the section between Yuanping West Station and Xinzhou West Station on the Datong-Xi'an High-Speed ​​Passenger Dedicated Line, and the total test data collection time was 21,000 seconds.

[0031] Preferably, the data from a single trip in the line test (including three typical operating conditions: deceleration, constant speed, and acceleration) is extracted according to the time-stress relationship, and the characteristics are studied in the time domain and frequency domain respectively, so as to intuitively show the dynamic stress changes of the bogie frame in the single trip test. Figures 4 to 12 The figure shows the dynamic stress signal waveforms and time-frequency diagrams at each measuring point. It can be seen that the dynamic stress change trends at each measuring point of the structure differ during operation. Figure 6 The test results show that there are several dynamic stress waveform fluctuations at the gearbox hanger test point (1-25) at the beginning of the test. This is caused by the gearbox hanger force due to the change in gearbox speed under the condition of train starting and accelerating. Therefore, the amplitude change is more obvious at low frequency. Figure 7The test point (1-29) on the brake jack shows obvious stress signal fluctuations at the end of the test. This is because the brake caliper generates braking force when the train enters the station, and the entire process is basically in a low amplitude state, indicating that the vibration at this part is relatively weak. Figure 9 The amplitude of the positioning arm measuring point (1-51) is greater than that of the other measuring points, but the overall trend does not show significant fluctuations.

[0032] Fatigue damage is a core parameter for assessing fatigue failure of mechanical components. Currently, the most commonly used damage assessment method in engineering projects is the Miner linear cumulative damage model. However, in most cases, this method is relatively conservative in assessing fatigue damage, using only simple linear superposition and failing to consider the nonlinear interaction effects of locomotive luffing impact loads. To compensate for this deficiency, this embodiment uses the PSN (probabilistic stress-life fatigue curve) curve of heavy-duty locomotive frame materials as a basis, combined with a modified Corten-Dolan nonlinear cumulative damage model, to analyze the cumulative damage at various measuring points of the bogie frame. Based on this, the life prediction of the bogie frame is analyzed.

[0033] Preferably, for the determination of PSN curves for heavy-duty locomotive frame materials, due to the dispersion of fatigue life, the conventional SN curve represents the median SN curve (S is the stress level, N is the fatigue life) with a survival rate P=50% and a confidence level ξ=50%. It can only reflect the relationship between the median fatigue life and various stress levels. Therefore, using the conventional SN curve to estimate the fatigue results of the structure often leads to a risky outcome. If a PSN curve with higher survival rate and confidence level is used to evaluate the structure, the results will be more inclined towards safety and reliability.

[0034] The PSN curve (probabilistic stress-life fatigue curve) of the heavy-haul locomotive frame material is double-logarithmic, and the mathematical expression is as follows: (1) In the formula, Let P be the fatigue life corresponding to a given survival rate P. This refers to the stress amplitude. , The undetermined coefficients in the PSN curve are calculated using the following formula: (2) (3) In the formula, Representing the Level stress amplitude; Representing the The fatigue life value corresponding to the stress amplitude; The th in logarithmic coordinates The value of the first-order stress; The th in logarithmic coordinates The fatigue life value corresponding to the stress level; is the stress level.

[0035] When determining the PSN curve, calculations are generally based on the group method and the rise and fall method. The fatigue life in the low stress level region of the curve can be determined by the rise and fall method, while the fatigue life in the high stress level region of the curve (generally level 3-4) is determined by the group method. In this fatigue test, two materials used for bogie frames were selected, with grades QA and SMA, respectively. The welding structures were corner joints and butt joints, respectively. The effective number of specimens was selected as 24. The fatigue test was conducted using these materials. After substituting the obtained data into equations (2) and (3), the calculation results are shown in Tables 1 and 2.

[0036] Table 1 shows the PSN curves for two welding structures using grade QA. Table 2 shows the PSN curves for two welding structures using grade SMA. Preferably, when correcting the nonlinear damage model, under variable amplitude loading, the sum of the damage increments caused by stresses at each level should be equal to a certain constant amplitude stress level. The damage caused by the action until the part fails is equal. Based on this principle, equation (4) is proposed to predict the fatigue life under multi-level stress loading: (4) In the formula, Total fatigue life; and These represent the maximum stress and its corresponding fatigue life in a series of loads applied to the specimen; For the first The proportion of cycles corresponding to the stress level to the total number of cycles; The stress level; This represents the material constant.

[0037] In fact, the index The fixed value of the exponent often leads to significant discrepancies between the predicted fatigue results and the actual results, resulting in a biased and unsafe phenomenon. Based on the correlation between fatigue damage and applied stress as discussed in the binary fatigue failure theory, and combined with the relationship between the exponent and applied load in the Corten-Dolan nonlinear cumulative damage model, a bivariate function that simultaneously considers the cycle frequency and the loading stress ratio is defined to replace the exponent. Furthermore, in practical engineering, variable amplitude loads often exist in multiple stages. Therefore, the influencing factor of the load loading sequence is also taken into account, and the material constant d can be derived as shown in the following mathematical expression: (5) In the formula, This represents the actual number of cycles corresponding to each stress order; This represents the fatigue life corresponding to each stress level. The material parameters can be obtained by fitting the data from fatigue tests on the bogie base material using the least squares method in conjunction with relevant fatigue failure criteria.

[0038] Combining the above formulas, a modified Corten-Dolan model (hereinafter referred to as the CD model) is obtained and introduced into the prediction of fatigue in high-speed train bogies. The modified CD model is expressed as follows: (6) In the formula, This indicates fatigue damage to the structure.

[0039] Preferably, in this embodiment, the operating mileage is selected as the service parameter, and a service parameter-damage regression equation is constructed by comparing the test mileage at each measuring point with the fatigue damage value. The method for establishing the service parameter-damage regression equation is as follows: 1) The measured data of the bogie frame for a single trip were divided into W segments using an equal-interval method. After extensive grouping calculations and considering various factors, the dynamic stress data of each measuring point were preferably grouped at t=100s. A one-dimensional stress spectrum was compiled for each segment, and combined with the modified CD model and the PSN curve of the heavy-duty locomotive frame material, the corresponding fatigue damage value was calculated. This value was then statistically compared with the mileage of each measuring point within the corresponding interval. Finally, a scatter plot of the test mileage and damage at each measuring point was obtained, as shown below. Figure 13 As shown.

[0040] 2) Based on the orthogonal distance regression algorithm, the mileage and fatigue damage value of each measuring point are nonlinearly fitted to obtain the service parameter-damage regression equation. The specific implementation process is as follows: Depend on Figure 13 It can be seen that the mileage at each measuring point and the damage have an approximately nonlinear relationship, and the trends are basically consistent. Therefore, this embodiment uses an orthogonal distance regression algorithm to perform nonlinear fitting on the calculated mileage and damage values ​​at each measuring point, which can simultaneously consider the error between the service parameters as independent variables x and the piecewise cumulative fatigue damage as dependent variable y. To simplify the complexity of the problem, the original data is mapped to a high-dimensional space, so that the data in the high-dimensional space are approximately linearly related and then calculated. Orthogonal regression uses the sum of squares of the horizontal and vertical coordinate residuals as the objective function to obtain the optimal solution, and its derivation process is as follows.

[0041] Assume the equation of the line to be fitted is: (7) In the formula, The coordinates of a point through which the line passes; Let be the normal vector of the line, which satisfies .

[0042] No. The distance from a point to a line can be represented as a vector. exist The length of the projection in the direction, hence the objective function It can be represented as: (8) In the formula, Indicates the first Points The distance to the fitted line.

[0043] Next, the objective function Find out about and Taking the partial derivative and setting it to 0, we get: (9) Dividing both sides of the above equation by the number of independent variables, n, we get: (10) In the formula, Let x be the mean, expressed as ; Let y be the mean, expressed as .

[0044] Obviously, point Since it satisfies the equation of a straight line, it must lie on the fitted line. Therefore, we can let... At this point, the objective function becomes: (11) Dividing the objective function by the number of independent variables n, we get: (12) In the formula, and Let x and y be the variances, respectively. Let x and y be the covariances, respectively. and For the correlation matrix, , ; For matrix The transpose of .

[0045] Therefore, this is a problem of finding the minimum value of a quadratic form. Because Since it is a symmetric matrix, it can be orthogonally diagonalized: (13) In the formula, and For matrix eigenvalues; and The corresponding feature vector; and For matrix and transpose; A matrix composed of eigenvectors; For matrix transpose; It is a diagonal matrix composed of eigenvalues.

[0046] Then we have: (14) make , , ,but: (15) In the formula, This is the intermediate column vector in the derivation process; and This is an intermediate quantity in the derivation process.

[0047] because It is an orthogonal matrix that satisfies ( (where it is the identity matrix), therefore Therefore, it can be concluded that For a unit vector (i.e., its magnitude is 1), satisfying .set up Therefore, it can be deduced that: when hour, Minimum value can be obtained .because ,and and All are unit vectors, therefore Therefore, the final result is: the normal vector of the fitted line. Equal to the corresponding matrix Minimum eigenvalue unit eigenvectors .

[0048] Based on the orthogonal distance regression algorithm, the linear relationship between the operating mileage and damage at each measuring point was fitted, as shown in Table 3. The expression for the service parameter-damage regression equation is: ,in This indicates the service parameters (in this embodiment, the operating mileage of each measuring point). The regression coefficients are determined by fitting the measured damage-mileage data using an orthogonal distance regression algorithm. The R-value is used as the evaluation index for the goodness of fit of the regression model. 2 Adjusted R 2 The residual sum of squares (RSS) is used as an evaluation metric to measure the error between the fitted values ​​of the regression model and the true values, thus assessing the fitting effect of the regression algorithm.

[0049] Table 3. Fitting results of linear equation parameters between mileage and damage at each measuring point As shown in the table, the nonlinear relationship between the mileage at each measuring point and the damage is highly consistent. Evaluation index R 2 Adjusted R 2 The value of is between 0 and 1. The closer the value is to 0, the worse the fit of the regression equation to the observed values; conversely, the better the fit. Both evaluation indices for all measurement points are above 0.95, indicating that the results fitted by the orthogonal distance regression algorithm are accurate. The evaluation index, Residual Sum of Squares (RSS), measures the error between the observed values ​​and the model's predicted values. The closer the value is to 1, the larger the fitting error; conversely, the smaller the error. The measurement point with the largest RSS value is the one installed at the swivel arm of the positioning seat (1-51), with a value of 2.77E-19, which is close to 0. This proves that the above fitting results are good, but further verification using hypothesis testing is needed.

[0050] After fitting the data to a curve using a regression algorithm, a significance test is needed to further verify the correctness of the fit, i.e., at a significance level of 100%. (This embodiment takes) This test examines whether there is a significant relationship between the independent variable x (i.e., service parameters) and the dependent variable y (i.e., segmented cumulative fatigue damage), and verifies the validity of the hypothesis. (16) In the formula, As the initial assumption; This is an opposing hypothesis; This represents the probability of the time to be tested occurring.

[0051] If accepted If the value is equal to 0, it means there is no significant relationship between the independent variable x and the dependent variable y; otherwise, it means there is a significant relationship between them.

[0052] In this embodiment, the F-test is selected to test the above hypothesis. Combining the F-statistic and quantiles, we can obtain: (17) In the formula, and These represent the maximum and minimum variances in the sample, respectively. This represents the number of samples.

[0053] The corresponding F-value can be calculated based on the sample values ​​for comparative testing. That is, not accepting the initial hypothesis. This proves that the above hypothesis fits; otherwise, it means accepting the initial hypothesis. This proves that the above assumptions do not fit.

[0054] Table 4 shows the significance test results of the hypothesis distribution for each measurement point (the probability values ​​in this table are notated using the exponential E notation, 7.66E-10, which is 7.66×10). -10 As shown in Table 4, all measuring points All statistical values ​​are greater than the significance level of 0.05. The test value is 4.96, and the minimum F-statistic is 172.38, corresponding to the measuring point (1-25) installed on the gearbox hanger. The minimum F-statistic is also much greater than... The basic idea of ​​hypothesis testing is to derive statistical judgments based on the principle of "low-probability events." As shown in Table 4, the probability of the hypothesis not fitting for each measurement point is close to 0. In summary, the F-test indicates that the hypothesis fit for the above measurement points is valid.

[0055] Table 4. Significance test results of the hypothesis distribution at each measurement point. For the prediction and verification of bogie frame life, preferably, the service parameters-damage regression equation fitted at the above-mentioned measuring points is substituted into the total track length of 672,500m to obtain the following results: Figure 14 The image shows a comparison of damage radar readings at various measuring points on the bogie frame calculated using different evaluation methods. It can be seen that among the nine measuring points, the fatigue damage value is highest at the measuring point installed on the positioning arm (1-51), while the fatigue damage value is lowest at the measuring point at the shock absorber installed on the crossbeam (1-12). The service parameter-damage regression equation (i.e., ...) is used... Figure 13The cumulative damage calculated by the diagram (shown as "nonlinear prediction") and the CD model is less than the linear cumulative damage of the Miner model, indicating that the fatigue results calculated by the traditional Miner linear cumulative damage model are relatively conservative. The cumulative damage values ​​calculated by the service parameter-damage regression equation are all greater than the damage values ​​calculated by the CD model. However, the trend of the cumulative damage values ​​calculated by the three evaluation methods at each measuring point is basically the same, indicating that the service parameter-damage regression equation is better than the CD model for fatigue evaluation of the bogie frame, and also proving that the method of evaluating the frame based on the regression equation is reliable.

[0056] It should be noted that when designing the lifespan of high-speed trains, a certain safety margin is generally reserved to ensure the reliability and safety of the train throughout its entire life cycle. This margin is typically 20%. Therefore, in this embodiment, the fatigue threshold of the bogie frame is preferably set to 0.8, and this fatigue threshold is substituted into the above-mentioned mileage-damage regression equation to calculate the mileage of the bogie frame throughout its entire life cycle, thus predicting the lifespan of the bogie frame.

[0057] Therefore, to reflect the final evaluation results, the Basquin fatigue equation SN curve data from the British standard BS 7608:2014 "Guide to Fatigue Design and Evaluation of Steel Structures" are selected as shown in Table 5. The BS 7608 standard classifies different fatigue levels according to the construction form of the welded joint; this embodiment uses two fatigue levels, F2 and G. The parameter m is the power exponent of the SN curve, and the parameter C is the SN curve fitting constant.

[0058] Table 5 Material parameters for damage calculation at each measuring point under BS 7608 standard The same method was used to calculate its mileage-damage regression equation, and it was compared with the lifespan predicted using the PSN curve. The results are as follows: Figure 15 As shown. Among them, the predicted life of the positioning arm measuring point (1-51) is the smallest among all measuring points, and the fatigue life of the entire bogie structure depends on the life of the weakest measuring point. Based on the BS standard parameters and PSN curve parameters, the predicted life of the bogie frame is 8.91 × 10⁻⁶. 9 m and 2.09×10 10 m. It can be seen that the train life predicted based on the PSN curve parameters is 2.09 × 10⁻⁶. 10 m is greater than the design life of 1.2 × 10⁻⁶ for Chinese high-speed trains. 10The value of m indicates that the calculation results of life prediction based on nonlinear regression equations are consistent with the fatigue life safety margin ensured by the bogie frame design, meaning that the method is reliable; the materials used in the aforementioned bogie frame can fully meet the design standard requirements. The train life predicted using PSN curve parameters is greater than the life calculated based on BS standard parameters, and the train life predicted using BS standard parameters is less than the train's design life, indicating that the life calculation results based on BS standard parameters are too conservative. Furthermore, the predicted life change trends at each measuring point are basically consistent under both parameters, and opposite to the equivalent stress change trends at each measuring point.

[0059] Based on the determined bogie frame fatigue threshold of 0.8, let the expression of the service parameter-damage regression equation be... The value is equal to the fatigue threshold, and the total lifespan mileage is solved by inverse calculation. (Take a physically valid positive root) Calculate the remaining fatigue life ;in, This indicates the locomotive's current cumulative mileage.

[0060] Preferably, in this embodiment, the remaining fatigue life result calculated in step four is connected to the locomotive digital twin operation and maintenance system. When the remaining life is lower than the preset safety threshold, the system automatically generates the maintenance priority ranking of the corresponding components and pushes the locomotive operation speed limit control suggestion to the dispatching platform to slow down the accumulation of fatigue damage by reducing the operating speed.

[0061] Example 2 Embodiment 2 of the present invention provides a heavy-haul locomotive life prediction system based on regression equations, corresponding to Embodiment 1 above, to implement the steps of the method in Embodiment 1 above, comprising: The data acquisition module is used to collect measured alternating stress time-series data of the load-bearing components of heavy-duty locomotives; The damage model construction module is used to construct a modified Corten-Dolan nonlinear cumulative damage model with load interaction correction terms, using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation and the measured alternating stress time series data collected by the data acquisition module as input. The regression fitting module is used to calculate fatigue damage at different service stages based on the nonlinear cumulative damage model and fit the service parameter-damage regression equation. The life prediction module is used to calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

[0062] The specific functions of each module described above are explained in the relevant content of the method in Embodiment 1, and will not be repeated here.

[0063] Example 3 Embodiment 3 of the present invention provides a terminal device corresponding to Embodiment 1 above, including a memory, a processor, and a computer program or instructions stored in the memory. When the processor executes the computer program or instructions, it implements the steps of the method of Embodiment 1 above.

[0064] Example 4 Embodiment 4 of the present invention provides a computer-readable storage medium corresponding to Embodiment 1 above, on which a computer program / instructions are stored. When the computer program / instructions are executed by a processor, they implement the steps of the method of Embodiment 1 above.

[0065] Example 5 Embodiment 5 of the present invention provides a computer program product corresponding to Embodiment 1 above, including a computer program or instructions, which, when executed by a processor, implements the steps of the method of Embodiment 1 above.

[0066] A computer-readable storage medium can be a tangible device that holds and stores instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof.

[0067] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0068] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0070] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0071] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for predicting the lifespan of heavy-haul locomotives based on regression equations, characterized in that, Includes the following steps: S1. Collect measured alternating stress time series data of the load-bearing components of heavy-duty locomotives; S2. Using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation, and using the measured alternating stress time series data collected in step S1 as input, a modified Corten-Dolan nonlinear cumulative damage model with load interaction correction term is constructed. The expression for the nonlinear cumulative damage model is as follows: ;in, This indicates fatigue damage to the load-bearing components of a heavy-haul locomotive. This represents the actual number of cycles corresponding to each stress order. This represents the fatigue life corresponding to each stress level. Representing the Level stress amplitude, and These represent the maximum stress and corresponding fatigue life among a series of loads applied to the load-bearing components of a heavy-duty locomotive. For stress level, These are material parameters; the load interaction correction term is used to characterize the accelerated deterioration effect of high-amplitude alternating loads on component fatigue; S3. Based on the nonlinear cumulative damage model, calculate the fatigue damage at different service stages and fit the service parameter-damage regression equation. S4. Calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

2. The method for predicting the lifespan of heavy-haul locomotives based on regression equations according to claim 1, characterized in that, The specific implementation process of step S3 includes: dividing the entire service range of the component, segmenting the alternating stress time series data according to the service range, inputting the alternating stress data of each segment into the nonlinear cumulative damage model, calculating the cumulative fatigue damage corresponding to each segment, and fitting the service parameter-damage regression equation with the service parameter as the independent variable and the segmented cumulative fatigue damage as the dependent variable.

3. The method for predicting the lifespan of heavy-haul locomotives based on regression equations according to claim 1, characterized in that, Service parameters-damage regression equation The expression is: ;in, Indicates service parameters, , , The regression coefficient is determined by fitting the measured damage-service parameter data using an orthogonal distance regression algorithm.

4. The method for predicting the lifespan of heavy-haul locomotives based on regression equations according to claim 3, characterized in that, Remaining fatigue life The expression is: ;in, The total service life mileage is obtained by inversely solving the service parameter-damage regression equation by setting the value of the service parameter to the fatigue threshold: , The fatigue threshold, This indicates the locomotive's current cumulative mileage.

5. The method for predicting the lifespan of heavy-haul locomotives based on regression equations according to claim 1, characterized in that, Based on the remaining fatigue life of the components calculated in step S4, maintenance guidance information for heavy-haul locomotives is generated. The maintenance guidance information includes at least one of component maintenance priority and locomotive speed limit control instructions.

6. A heavy-haul locomotive life prediction system based on regression equations, characterized in that, include: The data acquisition module is used to collect measured alternating stress time-series data of the load-bearing components of heavy-duty locomotives; The damage model construction module is used to construct a modified Corten-Dolan nonlinear cumulative damage model with load interaction correction terms, using the probabilistic PSN fatigue curve as the benchmark for single-cycle fatigue life calculation and the measured alternating stress time series data collected by the data acquisition module as input. The expression for the nonlinear cumulative damage model is as follows: ;in, This indicates fatigue damage to the load-bearing components of a heavy-haul locomotive. This represents the actual number of cycles corresponding to each stress order. This represents the fatigue life corresponding to each stress level. Representing the Level stress amplitude, and These represent the maximum stress and corresponding fatigue life among a series of loads applied to the load-bearing components of a heavy-duty locomotive. For stress level, These are material parameters; the load interaction correction term is used to characterize the accelerated deterioration effect of high-amplitude alternating loads on component fatigue; The regression fitting module is used to calculate fatigue damage at different service stages based on the nonlinear cumulative damage model and fit the service parameter-damage regression equation. The life prediction module is used to calculate the remaining fatigue life of the component based on the service parameters-damage regression equation.

7. A terminal device, comprising a memory, a processor, and a computer program stored in the memory; characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program / instructions stored thereon; characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program / instructions; characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method as described in any one of claims 1 to 5.

Images