Method, device and equipment for optimizing powertrain mounting system and storage medium

Abstract

This application relates to a method, apparatus, device, and storage medium for optimizing a powertrain mounting system. The method includes: acquiring a load report associated with the powertrain mounting system; obtaining load characteristic values ​​of the powertrain mounting system under various preset operating conditions from the load report; and obtaining the predicted driving time of a vehicle equipped with the powertrain mounting system under each preset operating condition; acquiring a load distribution model corresponding to each preset operating condition based on the load characteristic values; generating a load sequence corresponding to each preset operating condition based on the load distribution model and the predicted driving time; concatenating the load sequences corresponding to each preset operating condition to generate an extrapolated load-time history; and generating an extrapolated load spectrum based on the extrapolated load-time history; and performing robust optimization processing on the powertrain mounting system using the extrapolated load spectrum. This method enables accurate optimization of the powertrain mounting system using operating condition load reports.

Description

Technical Field

[0001] This application relates to the field of automotive technology, and in particular to a method, apparatus, computer device, computer-readable storage medium, and computer program product for optimizing a powertrain mounting system. Background Technology

[0002] With the development of automotive technology, a technology for optimizing the design of powertrain mounting systems has emerged. Powertrain mounting systems are important components of automobiles, used to support the powertrain and isolate vibrations. The durability and reliability of powertrain mounting systems directly affect the service life of automobiles as well as noise, vibration, and harshness (NVH) performance. Therefore, the optimized design of powertrain mounting systems is crucial in automotive engineering.

[0003] In traditional technologies, optimizing powertrain mounting system design relies on a complete load spectrum. However, due to limitations in data acquisition and cost, current load reports for powertrain mounting systems typically only record load characteristic values ​​under various preset operating conditions and the mileage percentage for each typical operating condition, lacking a complete load spectrum. Therefore, it is difficult to accurately optimize the powertrain mounting system using load reports alone. Summary of the Invention

[0004] Based on this, this application addresses the aforementioned technical problems by providing a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for optimizing a powertrain mounting system that can accurately optimize the powertrain mounting system.

[0005] In a first aspect, this application provides an optimization method for a powertrain mounting system, including:

[0006] Obtain the load report associated with the powertrain mounting system to be optimized, obtain the load characteristic value of the powertrain mounting system under each preset working condition from the load report, and obtain the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset working condition.

[0007] Based on the load characteristic values, obtain the load distribution model corresponding to each preset working condition, and generate the load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition.

[0008] By splicing the load sequences corresponding to each of the preset working conditions, an extrapolated load-time history associated with the powertrain mounting system is generated, and based on the extrapolated load-time history, an extrapolated load spectrum associated with the powertrain mounting system is generated.

[0009] The powertrain mounting system is robustly optimized using the extrapolated load spectrum.

[0010] In one embodiment, the load distribution model includes: a normal distribution model, wherein each load characteristic value includes the peak load of the powertrain mounting system under each preset operating condition; the step of obtaining the load distribution model corresponding to each preset operating condition based on each load characteristic value includes: obtaining a pre-constructed coefficient of variation, and the correspondence between the peak load and the standard deviation and mean of the normal distribution model; the coefficient of variation is used to characterize the correlation between the standard deviation and the mean; for any preset operating condition, based on the peak load of the preset operating condition, the coefficient of variation, and the correspondence between the peak load and the standard deviation and the mean, obtaining the standard deviation and mean of the normal distribution model corresponding to the preset operating condition; and constructing the load distribution model of the preset operating condition based on the standard deviation and mean of the normal distribution model corresponding to the preset operating condition.

[0011] In one embodiment, generating a load sequence corresponding to each preset operating condition based on the load distribution model corresponding to each preset operating condition and the predicted driving time corresponding to each preset operating condition includes: for any preset operating condition, generating a Gaussian white noise sequence matching the predicted driving time based on the predicted driving time corresponding to the preset operating condition; performing linear transformation processing on the Gaussian white noise sequence using the standard deviation and mean of the load distribution model corresponding to the preset operating condition to obtain the original load sequence corresponding to the preset operating condition; inputting the original load sequence into a preset low-pass filter, and outputting the load sequence corresponding to the preset operating condition through the low-pass filter.

[0012] In one embodiment, the original load sequence is composed of original load values ​​corresponding to multiple sampling points; the step of outputting the load sequence corresponding to the preset operating condition through the low-pass filter includes: obtaining the filter coefficients corresponding to the low-pass filter and the original load value corresponding to the current sampling point; the current sampling point is any one of the multiple sampling points; if the current sampling point is the first sampling point of the multiple sampling points, the original load value corresponding to the current sampling point is used as the filtered load value of the current sampling point; if the current sampling point is not the first sampling point, the filtered load value of the previous sampling point is obtained, and the filtered load value of the previous sampling point and the original load value corresponding to the current sampling point are weighted using the filter coefficients to obtain the filtered load value of the current sampling point; the load sequence corresponding to the preset operating condition is constructed using the filtered load values ​​of each current sampling point.

[0013] In one embodiment, obtaining the predicted driving time of a vehicle equipped with the powertrain mounting system under each preset operating condition includes: obtaining the mileage percentage of each preset operating condition in the entire vehicle lifecycle from the load report, and obtaining the driving speed of the vehicle under each preset operating condition; obtaining the driving time percentage of each preset operating condition based on the mileage percentage of each preset operating condition and the driving speed of each preset operating condition; and obtaining the predicted driving time of each preset operating condition based on the predicted total driving time of the vehicle under each preset operating condition and the driving time percentage of each preset operating condition.

[0014] In one embodiment, generating the extrapolated load spectrum associated with the powertrain mounting system based on the extrapolated load-time history includes: performing rainflow counting on the extrapolated load-time history; extracting the complete load cycles contained in the extrapolated load-time history, and the cycle amplitude and cycle mean of each complete load cycle; obtaining the amplitude range of each cycle amplitude and the mean range of each cycle mean; dividing multiple amplitude-mean intervals according to the amplitude range and mean range; obtaining the cycle frequency matching each amplitude-mean interval based on the cycle amplitude and cycle mean of each complete load cycle; constructing a rainflow matrix according to the cycle frequency matching each amplitude-mean interval; and using the rainflow matrix as the extrapolated load spectrum.

[0015] In one embodiment, the robust optimization of the powertrain mounting system using the extrapolated load spectrum includes: using the extrapolated load spectrum to obtain fatigue life assessment values ​​of the powertrain mounting system under different combinations of design parameters; constructing a Kriging approximation model based on each combination of design parameters and each fatigue life assessment value; the Kriging approximation model being used to characterize the correlation between the combination of design parameters and the fatigue life assessment value; using the Kriging approximation model to obtain a first combination of design parameters corresponding to the powertrain mounting system when the fatigue life assessment value is optimal; if the first combination of design parameters satisfies a preset robustness condition, using the first combination of design parameters as the optimized design parameter combination of the powertrain mounting system; if the first combination of design parameters does not satisfy the preset robustness condition, using the Kriging approximation model to obtain a second combination of design parameters corresponding to the powertrain mounting system when the fatigue life assessment value is optimal and satisfies the preset robustness condition, and using the second combination of design parameters as the optimized design parameter combination of the powertrain mounting system.

[0016] In one embodiment, obtaining the fatigue life assessment value of the powertrain mounting system under different combinations of design parameters using the extrapolated load spectrum includes: for any combination of design parameters, constructing a finite element model of the powertrain mounting system corresponding to the design parameter combination, applying a unit load to the finite element model to perform static analysis, and outputting the static stress response result of the powertrain mounting system under the design parameter combination; and generating the fatigue life assessment value of the powertrain mounting system under the design parameter combination based on the static stress response result and the extrapolated load spectrum.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0019] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in any of the above aspects.

[0020] The aforementioned optimization method, apparatus, computer equipment, computer-readable storage medium, and computer program product for the powertrain mounting system can obtain a load report associated with the powertrain mounting system to be optimized; extract load characteristic values ​​of the powertrain mounting system under various preset operating conditions from the load report; and obtain the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset operating condition. Based on each load characteristic value, a load distribution model corresponding to each preset operating condition is obtained; and based on the load distribution model and the predicted driving time, a load sequence corresponding to each preset operating condition is generated. The load sequences corresponding to each preset operating condition are then concatenated to generate an extrapolated load-time history associated with the powertrain mounting system; and based on the extrapolated load-time history, an extrapolated load spectrum associated with the powertrain mounting system is generated. The extrapolated load spectrum is then used to perform robust optimization processing on the powertrain mounting system. This application can extract load characteristic values ​​of the powertrain mounting system under various preset operating conditions from the load report of the powertrain mounting system, and obtain the predicted driving time of the vehicle corresponding to each preset operating condition. Thus, based on the load characteristic values, a load distribution model corresponding to each preset operating condition can be obtained. Combined with the predicted driving time corresponding to each preset operating condition, a load sequence corresponding to each preset operating condition can be generated, thereby extrapolating the load-time history. Based on the extrapolated load-time history, an extrapolated load spectrum can be generated, and robust optimization can be performed using the extrapolated load spectrum. In other words, this application can extrapolate the complete load spectrum based on the operating condition load report of the powertrain mounting system, and therefore the powertrain mounting system can be accurately optimized through the operating condition load report. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic flowchart of an optional method for optimizing a powertrain mounting system in one embodiment;

[0023] Figure 2 This is a schematic diagram of an optional process for obtaining a load distribution model in one embodiment;

[0024] Figure 3 This is a schematic diagram of an optional process for generating a payload sequence in one embodiment;

[0025] Figure 4 This is a schematic diagram of an alternative process for generating a payload sequence in another embodiment;

[0026] Figure 5 This is a schematic diagram of an optional process for generating an extrapolated load spectrum in one embodiment;

[0027] Figure 6 This is a schematic diagram of an optional process for robustness optimization in one embodiment;

[0028] Figure 7 This is a schematic diagram of an optional process for optimizing the fatigue life robustness of a powertrain mounting system in one embodiment;

[0029] Figure 8 A schematic diagram of an optional structure for an optimization device of the powertrain mounting system in one embodiment;

[0030] Figure 9 This is a schematic diagram of an optional internal structure of a computer device in one embodiment. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0032] The terms "first," "second," etc., used in this application may be used to describe various elements, but these elements are not limited by these terms. These terms are used only to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0033] In one embodiment, such as Figure 1 As shown, an optimization method for a powertrain mounting system is provided. This embodiment illustrates the method by applying it to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0034] Step S101: Obtain the load report associated with the powertrain mounting system to be optimized, obtain the load characteristic values ​​of the powertrain mounting system under each preset working condition from the load report, and obtain the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset working condition.

[0035] The powertrain mount system to be optimized refers to the powertrain mount system that needs robustness optimization. The load report is the operating condition load report for the powertrain mount system. This load report can be provided by the powertrain mount system supplier. The load report can record the load characteristic values ​​of the powertrain mount system under various preset operating conditions, such as the load characteristic values ​​of the powertrain mount system under various typical, extreme and misuse operating conditions. The predicted driving time refers to the predicted driving time of the vehicle equipped with the powertrain mount system under various typical, extreme and misuse operating conditions.

[0036] Specifically, when the terminal performs robustness optimization on the powertrain mounting system, it can first collect load reports provided by the powertrain mounting system supplier, and then extract the load characteristic values ​​of the powertrain mounting system under various preset operating conditions, namely, the powertrain mounting system under various typical, extreme and misuse operating conditions, from the load reports. It can also obtain the predicted driving time of the vehicle equipped with the above powertrain mounting system under each preset operating condition.

[0037] Step S102: Based on each load characteristic value, obtain the load distribution model corresponding to each preset working condition, and generate the load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition.

[0038] Load distribution models can be used to reflect the load distribution of the powertrain mounting system under various preset operating conditions. Generally, the load amplitude under each operating condition follows a certain probability distribution, and the probability distribution parameters of the load distribution model can be calibrated based on the load characteristic values ​​of each preset operating condition. Load sequences, on the other hand, refer to an ordered data sequence describing the changes in load on the powertrain mounting system over time. This sequence reflects the amplitude, sequence, and temporal characteristics of the load and is a key input for assessing structural fatigue life and performing damage accumulation analysis.

[0039] Specifically, after obtaining the load characteristic values ​​of the powertrain mounting system under various preset operating conditions, the terminal can use these load characteristic values ​​to calibrate the probability distribution parameters corresponding to the load distribution model under each operating condition, thereby obtaining the load distribution model corresponding to each preset operating condition. Subsequently, the terminal can combine the load distribution models corresponding to each preset operating condition with the predicted driving time to generate the load sequence corresponding to each preset operating condition.

[0040] Step S103: Assemble the load sequences corresponding to each preset working condition to generate the extrapolated load-time history associated with the powertrain mounting system, and generate the extrapolated load spectrum associated with the powertrain mounting system based on the extrapolated load-time history.

[0041] Extrapolated load-time history refers to the process of the load on the powertrain mounting system changing over time. This extrapolated load-time history can be obtained by splicing load sequences corresponding to each preset working condition. Extrapolated load spectrum refers to the complete load spectrum of the powertrain mounting system, which can be generated based on the extrapolated load-time history.

[0042] Specifically, after obtaining the load sequence corresponding to each preset working condition, the terminal can also splice the load sequences corresponding to each preset working condition to generate the extrapolated load-time history of the powertrain mounting system, thereby generating the extrapolated load spectrum associated with the powertrain mounting system using the extrapolated load-time history.

[0043] Step S104: The powertrain mounting system is robustly optimized using the extrapolated load spectrum.

[0044] Finally, after obtaining the complete extrapolated load spectrum, the powertrain mounting system can be robustly optimized using the extrapolated load spectrum to complete the optimization of the powertrain mounting system.

[0045] In the aforementioned optimization method for the powertrain mounting system, the following steps can be taken: First, obtain the load report associated with the powertrain mounting system to be optimized. Then, obtain the load characteristic values ​​of the powertrain mounting system under various preset operating conditions from the load report, and obtain the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset operating condition. Based on each load characteristic value, obtain the load distribution model corresponding to each preset operating condition. Then, based on the load distribution model and the predicted driving time, generate the load sequence corresponding to each preset operating condition. Finally, concatenate the load sequences corresponding to each preset operating condition to generate the extrapolated load-time history associated with the powertrain mounting system. Based on the extrapolated load-time history, generate the extrapolated load spectrum associated with the powertrain mounting system. Finally, use the extrapolated load spectrum to perform robust optimization processing on the powertrain mounting system. This application can extract load characteristic values ​​of the powertrain mounting system under various preset operating conditions from the load report of the powertrain mounting system, and obtain the predicted driving time of the vehicle corresponding to each preset operating condition. Thus, based on the load characteristic values, a load distribution model corresponding to each preset operating condition can be obtained. Combined with the predicted driving time corresponding to each preset operating condition, a load sequence corresponding to each preset operating condition can be generated, thereby extrapolating the load-time history. Based on the extrapolated load-time history, an extrapolated load spectrum can be generated, and robust optimization can be performed using the extrapolated load spectrum. In other words, this application can extrapolate the complete load spectrum based on the operating condition load report of the powertrain mounting system, and therefore the powertrain mounting system can be accurately optimized through the operating condition load report.

[0046] In one embodiment, the load distribution model includes: a normal distribution model, where each load characteristic value includes the peak load of the powertrain mounting system under various preset operating conditions; such as Figure 2 As shown, step S102 may further include:

[0047] Step S201: Obtain the pre-constructed coefficient of variation, and the correspondence between the peak load and the standard deviation and mean of the normal distribution model; the coefficient of variation is used to characterize the correlation between the standard deviation and the mean.

[0048] In this embodiment, it can be assumed that the load under each preset working condition follows a normal distribution. Therefore, the load distribution model used can be a normal distribution model, and the probability distribution parameters of the normal distribution model can include the mean. with standard deviation Meanwhile, the load characteristic values ​​used in this embodiment under each preset operating condition can refer to the peak load of the powertrain mounting system under each preset operating condition. For example, the operating condition report of the powertrain mounting system can record the peak values ​​of the three-dimensional forces Fx, Fy, Fz and the three-dimensional moments Mx, My, Mz at the mounting point under three preset operating conditions. The above peak values ​​can be used as the load characteristic values ​​under each preset operating condition.

[0049] The coefficient of variation is... This can refer to the correlation between the standard deviation and the mean, for example, it can characterize the ratio between the standard deviation and the mean, i.e. This ratio can be set to 0.2. The correspondence between the peak load and the standard deviation and mean of the normal distribution model refers to the calculation formula between the peak load and the standard deviation and mean of the normal distribution model. In this embodiment, the peak load corresponds to... Therefore, taking the peak value of the triaxial force Fx as an example, the relationship between the peak load and the standard deviation and mean of the normal distribution model can be expressed by the following formula:

[0050] Positive peak value:

[0051] Negative peak value:

[0052] Step S202: For any preset working condition, based on the load peak value, coefficient of variation, and the correspondence between the load peak value and the standard deviation and mean value, obtain the standard deviation and mean value of the normal distribution model corresponding to the preset working condition.

[0053] Step S203: Construct a load distribution model for the preset working condition based on the standard deviation and mean of the normal distribution model corresponding to the preset working condition.

[0054] After obtaining the above correspondence, the terminal can use the load peak value, coefficient of variation, and the correspondence between the load peak value and the standard deviation and mean of the normal distribution model under each preset working condition to calculate the standard deviation and mean of the normal distribution model under each preset working condition, and use the above standard deviation and mean to construct the normal distribution model under each preset working condition.

[0055] Taking the peak value of the triaxial force Fx as an example, the peak load values ​​under various working conditions are 1.2 KN, 8 KN, and -5 KN, respectively. Then, combining the coefficient of variation and the correspondence between the peak load value, standard deviation, and mean, the mean and standard deviation under each working condition can be calculated as shown in Table 1:

[0056] Table 1. Schematic diagram of mean and standard deviation under various operating conditions

[0057]

[0058] In this embodiment, the load distribution model can be a normal distribution model. The terminal can calculate the standard deviation and mean of the normal distribution model corresponding to each preset working condition based on the load peak value under each preset working condition, combined with the pre-constructed coefficient of variation, and the correspondence between the load peak value and the standard deviation and mean of the normal distribution model, so as to construct the load distribution model. This method can improve the accuracy of the load distribution model construction.

[0059] Furthermore, such as Figure 3 As shown, step S102 may further include:

[0060] Step S301: For any preset working condition, based on the predicted driving time corresponding to the preset working condition, generate a Gaussian white noise sequence that matches the predicted driving time.

[0061] Gaussian white noise sequences can be generated using the Monte Carlo method, and the number of sampling points in the generated Gaussian white noise sequence needs to match the predicted driving time corresponding to the preset driving conditions. For example, the number of sampling points in the Gaussian white noise sequence corresponding to the preset driving conditions can be obtained using the following expression:

[0062]

[0063] in, This can represent the number of sampling points in the Gaussian white noise sequence corresponding to any preset working condition i. This represents the predicted driving time corresponding to any preset driving condition i, while This can represent the sampling frequency, which can be set to 10 Hz.

[0064] For example, if the predicted driving time under each preset driving condition is 80h, 100h, and 120h, then the number of sampling points of the Gaussian white noise sequence corresponding to each preset driving condition are N1=80*3600*10=288000; N2=100*3600*10=360000; N3=120*3600*10=432000. For each driving condition i, an independent and identically distributed Gaussian white noise sequence is generated. This Gaussian white noise sequence follows a standard normal distribution with a mean of 0 and a standard deviation of 1. .

[0065] Step S302: Using the standard deviation and mean of the load distribution model corresponding to the preset working condition, the Gaussian white noise sequence is linearly transformed to obtain the original load sequence corresponding to the preset working condition.

[0066] The original load sequence corresponding to the preset working condition can be obtained by linearly transforming the Gaussian white noise sequence of the preset working condition using the standard deviation and mean of the load distribution model corresponding to the preset working condition. Specifically, it can be represented by the following formula:

[0067]

[0068] in, It can represent the original load sequence corresponding to the preset working condition i. This represents the mean of the load distribution model corresponding to the preset working condition i, while Let represent the standard deviation of the load distribution model corresponding to the preset working condition i. This represents the Gaussian white noise sequence corresponding to the preset operating condition i, and the generated noise sequence at this time. Follow the mean The standard deviation is The normal distribution is used as the original load sequence.

[0069] Step S303: Input the original load sequence into a preset low-pass filter, and output the load sequence corresponding to the preset working condition through the low-pass filter.

[0070] After obtaining the original load sequence, the original load sequence can be input into a preset low-pass filter. This preset low-pass filter can be a first-order low-pass filter. The original load sequence is filtered by the first-order low-pass filter to simulate the time correlation of the load, thereby obtaining the load sequence corresponding to the preset working condition.

[0071] In this embodiment, a Gaussian white noise sequence can be generated by predicting the driving time corresponding to the preset working conditions, and an original load sequence that follows a normal distribution with the corresponding standard deviation and mean can be generated using the standard deviation and mean. Finally, a low-pass filter is used for filtering to simulate time correlation. This method can improve the accuracy of load sequence generation.

[0072] Furthermore, the original load sequence consists of the original load values ​​corresponding to multiple sampling points; such as Figure 4 As shown, step S303 may further include:

[0073] Step S401: Obtain the filter coefficients corresponding to the low-pass filter and the original load value corresponding to the current sampling point; the current sampling point is any one of multiple sampling points.

[0074] In this embodiment, the original payload sequence can be composed of the original payload values ​​corresponding to multiple sampling points. The current sampling point refers to any one of the multiple sampling points. After the terminal inputs the original payload sequence into the low-pass filter, it can first obtain the filter coefficients corresponding to the low-pass filter and the original payload value corresponding to the current sampling point. The filter coefficients corresponding to the low-pass filter can be characterized by the following formula:

[0075]

[0076] In this embodiment 0.5Hz can be selected, but because Since the Hz frequency is set to 10 Hz, the filter coefficient can be set to 0.73.

[0077] Step S402: When the current sampling point is the first sampling point among multiple sampling points, the original load value corresponding to the current sampling point is used as the filtered load value of the current sampling point.

[0078] If the current sampling point is the first of multiple sampling points, then the original load value corresponding to the current sampling point can be directly used as the filtered load value of the current sampling point. That is, the filtered load value of the current sampling point can be expressed as: .

[0079] Step S403: If the current sampling point is not the first sampling point, obtain the filtered load value of the previous sampling point of the current sampling point, and use the filter coefficients to weight the filtered load value of the previous sampling point and the original load value corresponding to the current sampling point to obtain the filtered load value of the current sampling point.

[0080] If the current sampling point is not the first sampling point, the terminal can combine the filtered load value of the previous sampling point with the original load value corresponding to the current sampling point to obtain the filtered load value of the current sampling point. Specifically, the filtered load value of the previous sampling point and the original load value corresponding to the current sampling point are weighted using filter coefficients to obtain the filtered load value of the current sampling point. That is, the filtered load value of the current sampling point can be calculated using the following formula:

[0081]

[0082] in, This represents the filtered load value at the current sampling point t. This represents the filtered load value from the previous sampling point t-1. This represents the original load value at the current sampling point t, while This represents the filter coefficients. This filter does not change the mean, but reduces the variance and introduces time correlation.

[0083] Step S404: Construct a load sequence corresponding to the preset working condition using the filtered load values ​​of each current sampling point.

[0084] Finally, after obtaining the filtered load value of each current sampling point, the filtered load value of each current sampling point can be used to construct the load sequence corresponding to the preset working condition.

[0085] In this embodiment, the original load values ​​corresponding to each sampling point in the original load sequence can be filtered according to the filter coefficients corresponding to the low-pass filter, so that the filtered load sequence can introduce time correlation, thereby further improving the authenticity of the load sequence.

[0086] In one embodiment, step S101 may further include: obtaining the mileage percentage of each preset working condition in the entire life cycle of the vehicle from the load report, and obtaining the driving speed of the vehicle in each preset working condition; obtaining the driving time percentage of each preset working condition based on the mileage percentage of each preset working condition and the driving speed of each preset working condition; and obtaining the predicted driving time of each preset working condition based on the predicted total driving time of the vehicle in each preset working condition and the driving time percentage of each preset working condition.

[0087] In this embodiment, the load report can also record the mileage percentage of each preset working condition in the whole life cycle of the vehicle. In addition to obtaining the load characteristic value of the powertrain mounting system under each preset working condition from the load report, the terminal can also obtain the mileage percentage of each preset working condition in the whole life cycle of the vehicle. Then, the mileage percentage and the driving speed corresponding to each preset working condition can be used to calculate the driving time percentage corresponding to each preset working condition.

[0088] Specifically, the percentage of driving time in preset condition i can be calculated using the following formula:

[0089]

[0090] in, This indicates the percentage of driving time under preset condition i. This indicates the percentage of mileage in preset operating condition i. This represents the driving speed under preset condition i, while and These represent the mileage percentage and driving speed for each preset operating condition.

[0091] After obtaining the percentage of driving time corresponding to each preset driving condition, the predicted driving time for each preset driving condition can be calculated by combining the total predicted driving time of the vehicle under each preset driving condition, i.e., the total predicted driving time. For example, if the total predicted driving time of the vehicle is 300 hours, then the predicted driving time for each preset driving condition in the Fx channel can be shown in Table 2:

[0092] Table 2. Schematic diagram of predicted driving time under various operating conditions

[0093]

[0094] In this embodiment, the mileage percentage of each preset working condition in the whole life cycle of the vehicle can also be obtained from the load report, and the predicted driving time corresponding to each preset working condition can be calculated by using the mileage percentage. This method can improve the accuracy of obtaining the predicted driving time corresponding to each preset working condition.

[0095] In one embodiment, such as Figure 5 As shown, step S103 may further include:

[0096] Step S501: Perform rainflow counting on the extrapolated load-time history, extract the complete load cycles contained in the extrapolated load-time history, and the cycle amplitude and cycle mean of each complete load cycle.

[0097] Specifically, after obtaining the extrapolated load-time history, rainflow counting can be performed on the extrapolated load-time history using the ASTM E1049 standard to extract the complete load cycles contained in the extrapolated load-time history, as well as the cycle amplitude and cycle mean of each complete load cycle.

[0098] Step S502: Obtain the amplitude range of each cycle amplitude and the mean range of each cycle mean, and divide multiple amplitude-mean intervals according to the amplitude range and mean range.

[0099] Then, the terminal can first obtain the amplitude range of each cyclic amplitude and the mean range of the cyclic mean, and then divide the amplitude range and mean range at equal intervals to obtain multiple amplitude-mean intervals. For example, the amplitude range can be divided into 32 levels at equal intervals, and the mean can be divided into 32 levels at equal intervals, so that 32*32 amplitude-mean intervals can be obtained.

[0100] Continuing with the Fx channel as an example, since the overall load range is approximately... The comprehensive range estimates for each operating condition are shown in Table 3:

[0101] Table 3. Schematic diagram of load range for each working condition

[0102]

[0103] Therefore, the minimum value of the cyclic mean M is The maximum value of the cyclic mean M is The minimum value of the cyclic amplitude A is 0, and the maximum value of the cyclic amplitude A is:

[0104]

[0105] Therefore, the mean range of the cyclic mean is: cyclic mean M∈[-5,8], and the amplitude range of the cyclic amplitude is: cyclic amplitude A∈[0,6.5].

[0106] Since the cyclic mean and cyclic amplitude need to be divided into 32 levels respectively, the width of each level of the cyclic amplitude is... for:

[0107]

[0108] In order to enclose it, we can take The amplitude range of the j-th level: , j=1……,32.

[0109] Similarly, the width of each level of the cyclic mean for:

[0110]

[0111] In order to enclose it, we can take The mean interval of level i: , i=1……,32。

[0112] Step S503: Based on the cyclic amplitude and cyclic mean of each complete load cycle, obtain the cyclic frequency matching each amplitude-mean interval, construct the rainflow matrix according to the cyclic frequency matching each amplitude-mean interval, and use the rainflow matrix as the extrapolated load spectrum.

[0113] Finally, based on the rainflow counting results, the frequency of cycles falling within each amplitude-mean interval can be counted, constructing a 32×32 two-dimensional rainflow matrix for each load component. The row index of the matrix corresponds to the cyclic mean level, the column index corresponds to the cyclic amplitude level, and the matrix element value is the cyclic frequency of the load at that level.

[0114] Specifically, the two-dimensional rainflow matrix can be initialized by setting all element values ​​to 0. Then, the cyclic amplitude and cyclic mean of each complete load cycle can be traversed, and the amplitude-mean interval it satisfies can be determined based on the cyclic amplitude and cyclic mean. For example, if the cyclic amplitude of a certain complete load cycle satisfies the first-level amplitude interval and the cyclic mean satisfies the first-level mean interval, then the cyclic amplitude and cyclic mean of the complete load cycle fall into the amplitude-mean interval 1-1. At this time, the element values ​​of the two-dimensional rainflow matrix in [1,1] can be incremented by 1. After traversing in this way, the rainflow matrix can be constructed and used as the extrapolated load spectrum.

[0115] In this embodiment, the extrapolated load-time history can be processed by the rainflow counting method to obtain the extrapolated load spectrum associated with the powertrain mounting system, thereby improving the accuracy of the extrapolated load spectrum generation.

[0116] In one embodiment, such as Figure 6 As shown, step S104 may further include:

[0117] Step S601: Using the extrapolated load spectrum, obtain the fatigue life assessment value of the powertrain mounting system under different combinations of design parameters.

[0118] Different design parameter combinations refer to multiple pre-set combinations of design parameters. Fatigue life assessment values ​​refer to evaluation coefficients used to assess the fatigue life of the powertrain mounting system, such as fatigue damage coefficient, static strength safety factor, and mass coefficient. Specifically, after obtaining the extrapolated load spectrum, the terminal can input the extrapolated load spectrum into the Isight software, which then calculates the fatigue life assessment values ​​of the powertrain mounting system under various design parameter combinations.

[0119] Step S602: Construct a Kriging approximation model based on each combination of design parameters and each fatigue life assessment value; the Kriging approximation model is used to characterize the relationship between the combination of design parameters and the fatigue life assessment value.

[0120] The Kriging approximation model can be used to reflect the functional relationship between the combination of design parameters and the fatigue life assessment value. The model can be constructed based on each combination of design parameters and the corresponding fatigue life assessment value of the design parameter combination.

[0121] Specifically, after obtaining each combination of design parameters and the corresponding fatigue life assessment value, a Kriging model can be built in Isight's Approximation component, with the model set as follows: regression function: second-order polynomial, correlation function: Gaussian correlation function, parameter estimation: maximum likelihood estimation.

[0122] After constructing the Kriging approximation model, the accuracy of the constructed Kriging approximation model can be verified. If the accuracy verification is passed, it means that the generated Kriging approximation model can be used to characterize the correlation between the design parameter combination and the fatigue life assessment value. If the accuracy verification is not passed, a new design parameter combination needs to be added to obtain a new fatigue life assessment value, and the Kriging approximation model needs to be constructed again.

[0123] Step S603: Using the Kriging approximation model, obtain the first combination of design parameters for the powertrain mounting system under the condition of optimal fatigue life assessment value.

[0124] The first design parameter combination refers to the design parameter combination of the powertrain mounting system under the condition of optimal fatigue life assessment value. The first design parameter combination can be obtained by solving a multi-objective optimization problem on the Kriging approximation model.

[0125] Specifically, fatigue life assessment values ​​can include fatigue damage coefficient, static strength safety factor, and mass coefficient. Optimal fatigue life assessment value can refer to the optimization objective being to minimize the fatigue damage coefficient, maximize the minimum static strength safety factor, and minimize the mass coefficient. That is, using optimal fatigue life assessment value as the optimization objective, the Kriging approximation model is optimized and solved. The optimization solution can be completed using the multi-objective genetic algorithm NSGA-II. Finally, the first combination of design parameters corresponding to the powertrain mounting system under the condition of optimal fatigue life assessment value can be obtained.

[0126] Step S604: If the first design parameter combination meets the preset robustness conditions, the first design parameter combination is used as the optimized design parameter combination for the powertrain mounting system.

[0127] Due to the existence of processing errors, the first design parameter combination obtained at present may produce production errors in the real production environment. Therefore, after obtaining the first design parameter combination, the terminal can also perform a robustness assessment on the first design parameter combination, that is, determine whether the first design parameter combination meets the preset robustness conditions, and only when the first design parameter combination meets the preset robustness conditions will it be used as the optimized design parameter combination of the powertrain mounting system.

[0128] Specifically, the parameters in the first design parameter combination of the powertrain mounting system can be set to a random distribution. For example, the design parameter combination of the powertrain mounting system may include the mount wall thickness and the mount material. The mount wall thickness is set as a random variable following a normal distribution with a coefficient of variation of 0.012, while the mount material is set as a discrete distribution, which can be distinguished by material type. Then, a sub-sampling analysis can be performed using the Monte Carlo simulation method to calculate the fatigue damage value for each sample, thereby obtaining the distribution of fatigue damage values. The distribution of fatigue damage values ​​can then be used to determine whether the set conditions are met. For example, the distribution of fatigue damage values ​​may include the mean and standard deviation of the fatigue loss values, which can then be used to determine whether the set conditions are met.

[0129] For example, after obtaining the mean and standard deviation of fatigue loss values, the Sigma level can be calculated first, and then it can be determined whether the Sigma level meets the requirements. Quality requirements must be met at the Sigma level. Only when quality requirements are met is it stated that the first combination of design parameters satisfies the preset robustness conditions. The Sigma level can be calculated using the following formula: ,in This represents the Sigma level, indicating the minimum fatigue life required by the design specifications. It can be set to 1. This represents the mean of fatigue loss values, while The standard deviation of fatigue loss values ​​is only expressed when... Only under these circumstances can it be said that the Sigma level is satisfied. The quality requirement is that the first combination of design parameters must meet the preset robustness conditions.

[0130] Step S605: If the first design parameter combination does not meet the preset robustness conditions, the optimal fatigue life assessment value is obtained by using the Kriging approximation model, and the second design parameter combination corresponding to the powertrain mounting system is obtained under the preset robustness conditions. The second design parameter combination is then used as the optimized design parameter combination for the powertrain mounting system.

[0131] If the first combination of design parameters does not meet the preset robustness conditions, the Kriging approximation model needs to be optimized again using multiple objectives. This time, in addition to optimizing the fatigue life assessment value, the optimization objectives also need to ensure that the combination of design parameters meets the preset robustness conditions, i.e., the Sigma level must be satisfied. The quality requirement (Z≥6) is used as the optimization objective to solve the Kriging approximation model, thereby obtaining the second design parameter combination corresponding to the powertrain mounting system. The second design parameter combination is then used as the optimized design parameter combination for the powertrain mounting system.

[0132] In this embodiment, after obtaining the extrapolated load spectrum, the powertrain mounting system can be robustly optimized using the extrapolated load spectrum. This method transforms time-consuming simulation optimization into fast optimization on the Kriging approximation model, thereby reducing computational costs. Simultaneously, the optimization calculation process employs deterministic optimization... Level assessment and The three-step strategy of "robust optimization" first finds the performance boundary through multi-objective optimization, and then selects the design scheme that is not sensitive to fluctuations in input conditions through robustness analysis. This effectively solves the problem that the performance of traditional optimization design is prone to degradation in actual production applications, and significantly improves the engineering practicality and quality consistency of the product.

[0133] In one embodiment, step S601 may further include: for any combination of design parameters, constructing a finite element model of the powertrain mounting system under the combination of design parameters, applying a unit load to the finite element model to perform static analysis, and outputting the static stress response result of the powertrain mounting system under the combination of design parameters; and generating a fatigue life assessment value of the powertrain mounting system under the combination of design parameters based on the static stress response result and the extrapolated load spectrum.

[0134] Specifically, for any combination of design parameters, the terminal can construct a finite element model of the powertrain mounting system under that parameter combination. Abaqus software is then used to apply a unit load to the finite element model for static analysis, outputting the static stress response results of the powertrain mounting system under the design parameter combination. These results can be recorded in an ODB file. The terminal can then import the ODB file and the extrapolated load spectrum into Femfat software. Femfat automatically associates the unit load case in the ODB file with the corresponding rainflow matrix and uses the quasi-static superposition principle to linearly combine the load spectrum with the unit stress field, obtaining the stress history over time. Material SN curves are set, mean stress correction is selected, and Miner's linear cumulative damage theory is used. The fatigue solver is run to obtain the fatigue damage contour map of the mounting system, and the fatigue life assessment value under the design parameter combination is extracted.

[0135] Furthermore, all of the above processes can be integrated through Isight. A new optimization workflow can be created in Isight, adding two Simcode components. Simcode-1 is used to call Abaqus. Design parameter combinations (such as suspension wall thickness and suspension material type) are written into the corresponding locations in the Abaqus input file (.inp); the execution commands of the Abaqus solver are set; and the stress result file path is extracted from the Abaqus ODB file and passed to the next component. Simcode-2 is used to call Femfat. The extrapolated load spectrum file path and ODB file path are written into the Femfat input file (.fes); the execution commands of the Femfat solver are set; and fatigue life assessment values ​​are extracted from the Femfat result file using Excel software.

[0136] In this embodiment, the fatigue life assessment value can be automatically calculated under various combinations of design parameters, which can improve the efficiency of fatigue life assessment value calculation.

[0137] In one embodiment, a robust optimization design method for the fatigue life of a powertrain mounting system is also provided, such as... Figure 7 As shown, the method may include the following steps:

[0138] Step 1: Extrapolate the load spectrum based on the operating condition report.

[0139] 1.1 Obtaining Operating Load Data: Obtain the powertrain mounting system operating condition report provided by the supplier. This report includes the load characteristic values ​​of the three-dimensional forces Fx, Fy, Fz and the three-dimensional moments Mx, My, Mz at each mounting point under k typical durability conditions. Additionally, inquire with the supplier about the mileage percentage of each typical operating condition over the entire vehicle lifecycle (typical, misuse, and extreme operating conditions, a total of n operating conditions). Representative driving speeds for each typical operating condition Therefore, the time percentage of each typical operating condition in the entire vehicle life cycle (typical, misuse, extreme operating conditions) is as follows:

[0140]

[0141] The load characteristic value is the load value (peak value) corresponding to the maximum absolute value of the load under each working condition.

[0142] Taking the Fx channel as an example, the total duration of each working condition in this embodiment is 300 hours. Therefore, the load characteristic values ​​and driving time of each working condition can be shown in Table 2.

[0143] 1.2 Determine the load distribution model: Based on the random characteristics of the load on the suspension system, it is assumed that the load amplitude under each working condition follows a certain probability distribution, where the distribution parameters are calibrated according to the load characteristic values ​​provided by the supplier.

[0144] Specifically, it is assumed that the load under each operating condition follows a normal distribution, and the peak value corresponds to... Boundary (under normal distribution, since the peak corresponds to) The boundary, so the instantaneous value of the load falls within the boundary. The probability of the interval is as high as 99.73%, which can be considered as the actual load almost never exceeding the boundary, so we adopt... The boundary corresponds to the peak value (this can be changed if a higher confidence interval is required); the coefficient of variation in this embodiment is taken as... Standard deviation Corresponding positive peak value: Corresponding to negative peak value: Based on the above formulas, the standard deviation and mean of the normal distribution model corresponding to each operating condition under the Fx channel can be calculated, as shown in Table 1.

[0145] 1.3 Generating Extrapolated Load-Time Histories: Based on the load distribution model and duration proportion of each working condition, the extrapolated load-time histories for each working condition are randomly generated using the Monte Carlo method. During the generation process, the total time axis is divided into time periods corresponding to each working condition. Within each time period, load sequences are randomly generated according to the corresponding distribution model, and a filtered white noise method is introduced to simulate the time correlation of the loads.

[0146] Temporal correlation refers to the relationship between the load value at the current moment and the load value at the previous moment, rather than completely independent random jumps. Real loads (such as road loads and wind loads) have inertia and almost never change instantaneously. Filtering is precisely to simulate this physical characteristic. If white noise is used directly, although there are peaks and troughs, the changes are too drastic, producing a large number of "small jagged edges," which simply do not occur in the real world. Rainflow counting will record many meaningless small cycles, leading to distortion in damage calculations. After filtering, the jagged edges are smoothed out, leaving cycles that are closer to reality, and the rainflow counting results are more reliable.

[0147] Specifically, for each load component under each operating condition, a Gaussian white noise sequence is generated using the Monte Carlo method, and then filtered using a first-order low-pass filter (the cutoff frequency is set according to the actual load spectrum characteristics, e.g., 5Hz) to simulate the time correlation of the load. The generated sequence length (number of sampling points) is... for (Can be taken as needed) The sequences generated for each operating condition are spliced ​​together in the order of the operating conditions to obtain the Fx channel extrapolated load-time history with a total duration of 300 hours.

[0148] According to Table 2, the predicted driving time under each preset working condition is 80h, 100h, and 120h, respectively. Therefore, the number of sampling points for each working condition is N1 = 80 * 3600 * 10 = 288000; N2 = 100 * 3600 * 10 = 360000; N3 = 120 * 3600 * 10 = 432000. For each working condition i, an independent and identically distributed Gaussian white noise sequence is generated. , can be represented as After adding the mean and standard deviation, ,at this time, It can be generated using MATLAB or Python, but the sequence is white noise and has no time correlation.

[0149] A first-order low-pass filter (exponential smoothing) is used. Filter coefficients:

[0150]

[0151] According to the needs, 0.5Hz can be selected, therefore It is approximately 0.73.

[0152] Filtered load sequence:

[0153]

[0154]

[0155] Where t = 2, 3, 4...N i Filtering does not change the mean, but it reduces the variance and introduces time correlation.

[0156] For this embodiment, the sequence of each working condition is spliced ​​together:

[0157]

[0158] The total duration was 300 hours, with 10,800,000 sampling points.

[0159] 1.4 Rainflow Counting Processing: Rainflow counting is performed on the generated extrapolated load-time history to extract all complete load cycles, obtaining the cycle amplitude A and cycle mean M for each cycle, forming the original cycle dataset. Taking Fx as an example, the amplitude and mean of all complete cycles are statistically analyzed. Perform rainflow counting, extracting all complete cycle amplitudes A and cycle mean M. The results are compiled into a list, with each row containing (A...). k M k ).

[0160] 1.5. Equal-interval grading: Divide the load amplitude range into 32 equal-interval levels, and divide the load mean range into 32 equal-interval levels. Determine the upper and lower limits of the grading to ensure coverage of all load data. Specifically, the rainflow counting statistical results of the Fx channel can be divided into 32*32 equal-interval levels. Repeat this process for the other five channels. The overall load range is approximately... The comprehensive range estimates for each working condition are shown in Table 3.

[0161] Therefore, the minimum value of the cyclic mean M is The maximum value of the cyclic mean M is The minimum value of the cyclic amplitude A is 0, and the maximum value of the cyclic amplitude A is:

[0162]

[0163] Therefore, the mean range of the cyclic mean is: cyclic mean M∈[-5,8], and the amplitude range of the cyclic amplitude is: cyclic amplitude A∈[0,6.5].

[0164] Since the cyclic mean and cyclic amplitude need to be divided into 32 levels respectively, the width of each level of the cyclic amplitude is... for:

[0165]

[0166] In order to enclose it, we can take The amplitude range of the j-th level: , j=1……,32.

[0167] Similarly, the width of each level of the cyclic mean for:

[0168]

[0169] In order to enclose it, we can take The mean interval of level i: , i=1……,32。

[0170] 1.6 Constructing the Rainflow Matrix (Extrapolated Load Spectrum): Count the frequency of cycles falling within each amplitude-mean interval to construct an M×N two-dimensional rainflow matrix R. Matrix elements. This represents the frequency of cycles where the mean value is in the i-th interval and the amplitude value is in the j-th interval. This matrix fully describes the amplitude-mean-frequency distribution characteristics of the load and is the direct input for subsequent fatigue damage calculations (the representative values ​​of stress amplitude and average stress are taken as the median values ​​of each interval). For a six-component load, six rainflow matrices are constructed corresponding to Fx, Fy, Fz, Mx, My, and Mz. Specifically, a 32*32 zero matrix R can be initialized first. Traversing each rainflow cycle, the interval i containing the mean value and the interval j containing the amplitude value are determined, and let: The final matrix R is the rainflow matrix, which is the extrapolated load spectrum.

[0171] Step 2: Obtain ODB file using finite element analysis (unit load method).

[0172] 2.1 Establishing the Finite Element Model: A finite element model of the powertrain suspension system is established, including the suspension brackets, hybrid housing, engine, etc., and material properties are assigned. The suspension brackets and hybrid housing housing are meshed using second-order tetrahedral elements (C3D10M), while the bolts are meshed using C3D8 elements. Material properties are assigned: linear elastic material is used.

[0173] 2.2 Applying Unit Load Boundary Conditions: In Abaqus, apply unit loads in six directions at the mount point: Fx=1kN, Fy=1kN, Fz=1kN, Mx=1kN·m, My=1kN·m, and Mz=1kN·m, and perform six independent static analyses. Specifically, a reference point can be established at the engine connection point of the mount system, and kinematic coupling constraints can be established with the connection point surface. Full constraints are applied to the interface between the hybrid gearbox and the engine.

[0174] 2.3 Submit Calculation and Output ODB File: Run the Abaqus solver and output an ODB file containing stress influence coefficients for the six load directions. Specifically, submit the calculation in Abaqus / Standard, using the general statics analysis step. Set the output field variables such as stress and strain. After the calculation is complete, output an ODB file containing the load case and stress response results. This ODB file records the stress distribution of the suspension system under a unit load, i.e., the stress influence coefficient matrix.

[0175] Step 3: Femfat fatigue damage calculation.

[0176] Import the ODB file from step 2 and the six rainflow matrices from step 1 into the Femfat fatigue analysis software. Femfat uses the quasi-static superposition principle to linearly combine the unit stress field with the true load spectrum to obtain the true stress history: Among them, Fx(t) is the load sequence reconstructed from the rainflow matrix. Combining the material SN curve, the modified Hager plot (which equates the non-zero mean cyclic fatigue limit to the zero mean cyclic fatigue limit), and Miner's linear cumulative damage theory, fatigue damage and minimum fatigue life are calculated.

[0177] Step 4: Build an integrated automation process based on Isight.

[0178] By utilizing the Simcode component in Isight, Abaqus and Femfat are integrated into an automated workflow, achieving a fully automated loop of "design variable change → Abaqus solution → Femfat fatigue analysis → fatigue damage output". Specifically, a new optimization workflow is created in Isight, and two Simcode components are added:

[0179] Simcode-1: Invoke Abaqus. Write design variables (such as suspension wall thickness and suspension material type) into the corresponding locations in the Abaqus input file (.inp); set the execution command for the Abaqus solver; extract the stress result file path from the Abaqus ODB file and pass it to the next component.

[0180] Simcode-2: Invoking Femfat. Write the rainflow matrix file path from step 1 and the ODB file path from the previous step into the Femfat input file (.fes); set the Femfat solver to execute commands; use Excel software to extract the minimum fatigue damage value from the Femfat result file.

[0181] The two components are linked together using the Simcode data connector to form a closed loop: "Design variables → Abaqus solver → Femfat solver → Fatigue damage output". The parsing rules for the input and output files are configured to ensure correct updates in each loop.

[0182] Step 5: Deterministic multi-objective optimization based on optimized Latin hypercube sampling and Kriging model.

[0183] 5.1 Design of Experiment (DOE): An optimized Latin hypercube sampling method is used to generate initial sample points (combinations of design parameters) within the design space, ensuring that the sample points uniformly fill the design space. The number of sample points is at least 10 times the number of design variables; in this embodiment, 50 initial sample points are used.

[0184] 5.2 Call the integrated process to calculate the response: For each combination of design parameters, the integrated process in step 4 is automatically called to calculate the corresponding maximum fatigue damage value, minimum static strength safety factor and system mass (fatigue life assessment value), and to construct a sample dataset of design parameter combinations (design variables such as material, wall thickness, etc.) and outputs (fatigue damage, static strength safety factor, lightweight mass), resulting in a total of 50 sets of sample datasets (input design parameter combinations and output fatigue life assessment values).

[0185] 5.3 Constructing a Kriging Approximation Model: Based on the sample dataset, a Kriging model is used to construct an approximate functional relationship between design variables and fatigue damage, static strength safety factor, and lightweight mass. The Kriging model consists of a multinomial regression part and a stochastic process part, which can accurately interpolate sample points and provide an error estimate of the predicted values. Specifically, a Kriging model can be constructed based on 50 sample points in Isight's Approximation component. The model settings are as follows:

[0186] Regression function: Second-order polynomial;

[0187] Correlation function: Gaussian correlation function;

[0188] Parameter estimation: maximum likelihood estimation.

[0189] 5.4 Model Accuracy Validation: Randomly generate several validation points, compare the predicted values ​​of the Kriging model with the actual simulation values, and calculate the coefficient of determination R² and the root mean square error RMSE. If R² < 0.9 or RMSE exceeds the allowable threshold, supplement the sample points and update the model until the accuracy requirements are met.

[0190] 5.5 Multi-objective optimization based on approximation models: Set up a multi-objective optimization problem in Isight (the optimization objectives may be contradictory, which can be addressed using the NSGA-II multi-objective optimization algorithm):

[0191] Design variables: X1 (suspension wall thickness 15-25mm), X2 (suspension material type: ADC12, A280, SUS304).

[0192] Constraints: Minimum static strength safety factor ≥ 1.1, maximum fatigue damage < 1;

[0193] Optimization objectives: Minimize maximum fatigue damage, maximize minimum static strength safety factor, and minimize mass (lightweight design).

[0194] Optimization algorithm: NSGA-II, population size 100, number of generations 200.

[0195] The NSGA-II genetic algorithm is used to optimize the model on the Kriging approximation, which takes significantly less time than direct simulation. The Pareto front solution set is obtained, from which candidate deterministic optimization solutions are selected, denoted as solution A.

[0196] Step 6: Deterministic Optimization Solution Level assessment.

[0197] Due to manufacturing errors, the suspended wall thickness in the design variables is set as a random variable following a normal distribution, with a coefficient of variation of 0.012 (which can be adjusted according to actual needs). The suspended material is set as a discrete distribution, which can be differentiated according to material type. A Monte Carlo simulation method is used to perform 500 sampling analyses. The integrated workflow in step 4 is called to calculate output quantities such as fatigue damage distribution, and the mean fatigue loss value is obtained. and standard deviation Calculate the Sigma level based on the minimum fatigue damage threshold D required by the design specifications: Assess whether the Sigma level is met. Quality requirements (i.e., Z≥6).

[0198] Among them, Solution_A Analysis results: The mean fatigue damage of this deterministic solution Standard deviation The design specification requires a minimum fatigue life of D=1, and the calculated Sigma level is Z=2.74. Z is much less than 6, indicating that this deterministic solution has a high risk of failure under fluctuations in input variables, necessitating robust optimization.

[0199] Step 7: Robustness optimization.

[0200] If the Sigma level assessed in step 6 is not met If required, then start. Robustness optimization. Based on the Kriging approximation model constructed in step 5, the optimization aims to minimize the mean maximum fatigue damage, maximize the mean minimum static strength safety factor, minimize the mean mass, and satisfy... With a target level of 100 or higher, NSGA-II was used in conjunction with Monte Carlo simulation to find a suitable solution. The robust optimal solution for quality requirements.

[0201] Specifically, it can be started Robust optimization is also based on the Kriging approximation model constructed in step 5:

[0202] Design variables: Same as step 5, considering their distribution characteristics;

[0203] Optimization objectives: Minimize the mean maximum fatigue damage, maximize the mean minimum static strength safety factor, minimize the mean mass, and satisfy the following conditions: The target level is level 6 or above (Z≥6).

[0204] Optimization algorithm: NSGA-II combined with Monte Carlo simulation, performing 100 sampling evaluations on the design point in each iteration, and calculating... and .

[0205] After optimization and iteration, a new robust design point, solution_B, was obtained: wall thickness of 19.91 mm and material type ADC12. This combination of design parameters was then analyzed. All horizontal assessments met the requirements, with the average fatigue damage value being [missing information]. Standard deviation The Sigma level Z=6.78 satisfies the condition. Quality requirements. The design parameter combination B is the robust optimal solution output in this embodiment.

[0206] Fatigue life Prediction can be obtained from damage values: Focusing on this case, the total payload duration Robust optimal solution damage Therefore, lifespan prediction Fatigue life can also be converted into the number of cycles. In this embodiment, the test duration is used as the life prediction indicator.

[0207] This embodiment addresses the issue of missing data. Even when the supplier only provides limited load characteristic values, a rainflow matrix consistent with actual statistical characteristics is successfully constructed through load distribution extrapolation and pseudo-time history generation techniques. This provides an accurate input load spectrum for subsequent fatigue analysis, filling the technological gap in constructing a load spectrum from limited information. Furthermore, load simulation is more accurate: using the rainflow matrix to describe the amplitude-mean-frequency distribution of the load more realistically reflects the random load characteristics of actual working conditions compared to a single load value or simple waveform, improving the input accuracy of fatigue damage calculation. Simultaneously, fatigue calculation is more accurate: combining Abaqus finite element stress results (unit load method) with the rainflow matrix in the professional fatigue software Femfat, and employing the quasi-static superposition principle, multi-axis fatigue life prediction based on the actual load spectrum is achieved, with results more reliable than traditional empirical formula methods. The process is also automated; integrating Abaqus and Femfat through the Isight Simcode component enables a fully automated closed loop of design-analysis-optimization, significantly improving optimization efficiency and reducing manual intervention costs. And significantly improved optimization efficiency: By optimizing Latin hypercube sampling combined with the Kriging approximation model, time-consuming simulation optimization is transformed into fast optimization on the approximation model, reducing computational costs by more than 90%, making robust optimization feasible in engineering. Furthermore, the optimization results are reliable: employing "deterministic optimization, Level assessment and The three-step strategy of "robust optimization" first finds the performance boundary through multi-objective optimization, and then selects the design scheme that is not sensitive to fluctuations in input conditions through robustness analysis. This effectively solves the problem that the performance of traditional optimization design is prone to degradation in actual production applications, and significantly improves the engineering practicality and quality consistency of the product.

[0208] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0209] Based on the same inventive concept, this application also provides an optimization device for a powertrain mounting system to implement the optimization method for the powertrain mounting system described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the powertrain mounting system optimization device provided below can be found in the limitations of the powertrain mounting system optimization method described above, and will not be repeated here.

[0210] In one embodiment, such as Figure 8 As shown, an optimization device for a powertrain mounting system is provided, comprising: a load information acquisition module 801, a load sequence generation module 802, an extrapolated load spectrum construction module 803, and a robustness optimization module 804, wherein:

[0211] The load information acquisition module 801 is used to acquire the load report associated with the powertrain mounting system to be optimized, acquire the load characteristic value of the powertrain mounting system under each preset working condition from the load report, and acquire the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset working condition.

[0212] The load sequence generation module 802 is used to obtain the load distribution model corresponding to each preset working condition based on the load feature value, and generate the load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition.

[0213] The extrapolation load spectrum construction module 803 is used to splice the load sequences corresponding to each of the preset working conditions, generate the extrapolation load-time history associated with the powertrain mounting system, and generate the extrapolation load spectrum associated with the powertrain mounting system based on the extrapolation load-time history.

[0214] The robustness optimization module 804 is used to perform robustness optimization processing on the powertrain mounting system using the extrapolated load spectrum.

[0215] In one embodiment, the load distribution model includes: a normal distribution model, wherein each load characteristic value includes the peak load of the powertrain mounting system under each preset operating condition; a load sequence generation module 802, further configured to obtain a pre-constructed coefficient of variation, and the correspondence between the peak load and the standard deviation and mean of the normal distribution model; the coefficient of variation is used to characterize the correlation between the standard deviation and the mean; for any preset operating condition, based on the peak load of the preset operating condition, the coefficient of variation, and the correspondence between the peak load and the standard deviation and the mean, the standard deviation and mean of the normal distribution model corresponding to the preset operating condition are obtained; and a load distribution model for the preset operating condition is constructed based on the standard deviation and mean of the normal distribution model corresponding to the preset operating condition.

[0216] In one embodiment, the load sequence generation module 802 is further configured to, for any of the preset operating conditions, generate a Gaussian white noise sequence that matches the predicted driving time corresponding to the preset operating condition; perform linear transformation processing on the Gaussian white noise sequence using the standard deviation and mean of the load distribution model corresponding to the preset operating condition to obtain the original load sequence corresponding to the preset operating condition; input the original load sequence into a preset low-pass filter, and output the load sequence corresponding to the preset operating condition through the low-pass filter.

[0217] In one embodiment, the original load sequence consists of original load values ​​corresponding to multiple sampling points; the load sequence generation module 802 is further configured to: the current sampling point is any one of the multiple sampling points; when the current sampling point is the first sampling point of the multiple sampling points, use the original load value corresponding to the current sampling point as the filtered load value of the current sampling point; when the current sampling point is not the first sampling point, obtain the filtered load value of the previous sampling point of the current sampling point, and use the filter coefficients to weight the filtered load value of the previous sampling point and the original load value corresponding to the current sampling point to obtain the filtered load value of the current sampling point; and construct the load sequence corresponding to the preset working condition using the filtered load values ​​of each current sampling point.

[0218] In one embodiment, the load information acquisition module 801 is further configured to obtain the mileage percentage of each preset working condition in the entire life cycle of the vehicle from the load report, and obtain the driving speed of the vehicle in each preset working condition; obtain the driving time percentage of each preset working condition based on the mileage percentage of each preset working condition and the driving speed of each preset working condition; and obtain the predicted driving time of each preset working condition based on the predicted total driving time of the vehicle in each preset working condition and the driving time percentage of each preset working condition.

[0219] In one embodiment, the extrapolated load spectrum construction module 803 is further configured to perform rainflow counting on the extrapolated load-time history, extract the complete load cycles contained in the extrapolated load-time history, and the cycle amplitude and cycle mean of each complete load cycle; obtain the amplitude range of each cycle amplitude and the mean range of each cycle mean; divide multiple amplitude-mean intervals according to the amplitude range and mean range; obtain the cycle frequency matching each amplitude-mean interval based on the cycle amplitude and cycle mean of each complete load cycle; construct a rainflow matrix according to the cycle frequency matching each amplitude-mean interval; and use the rainflow matrix as the extrapolated load spectrum.

[0220] In one embodiment, the robustness optimization module 804 is further configured to: utilize the extrapolated load spectrum to obtain fatigue life assessment values ​​of the powertrain mounting system under different combinations of design parameters; construct a Kriging approximation model based on each combination of design parameters and each fatigue life assessment value; the Kriging approximation model characterizes the correlation between the combination of design parameters and the fatigue life assessment value; utilize the Kriging approximation model to obtain a first combination of design parameters corresponding to the powertrain mounting system when the fatigue life assessment value is optimal; if the first combination of design parameters satisfies a preset robustness condition, use the first combination of design parameters as the optimized combination of design parameters for the powertrain mounting system; if the first combination of design parameters does not satisfy the preset robustness condition, utilize the Kriging approximation model to obtain a second combination of design parameters corresponding to the powertrain mounting system when the fatigue life assessment value is optimal and satisfies the preset robustness condition, and use the second combination of design parameters as the optimized combination of design parameters for the powertrain mounting system.

[0221] In one embodiment, the robustness optimization module 804 is further configured to construct a finite element model of the powertrain mounting system under any combination of design parameters, apply a unit load to the finite element model to perform static analysis, and output the static stress response result of the powertrain mounting system under the combination of design parameters; and generate a fatigue life assessment value of the powertrain mounting system under the combination of design parameters based on the static stress response result and the extrapolated load spectrum.

[0222] Each module in the aforementioned powertrain mounting system optimization device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0223] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements an optimization method for a powertrain mounting system. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0224] Those skilled in the art will understand that Figure 9The structure shown is a block diagram of a partial structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0225] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0226] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0227] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0228] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0229] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program mentioned can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0230] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0231] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An optimization method for a powertrain mounting system, characterized in that, The method includes: Obtain the load report associated with the powertrain mounting system to be optimized, obtain the load characteristic value of the powertrain mounting system under each preset working condition from the load report, and obtain the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset working condition. Based on the load characteristic values, obtain the load distribution model corresponding to each preset working condition, and generate the load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition. By splicing the load sequences corresponding to each of the preset working conditions, an extrapolated load-time history associated with the powertrain mounting system is generated, and based on the extrapolated load-time history, an extrapolated load spectrum associated with the powertrain mounting system is generated. The powertrain mounting system is robustly optimized using the extrapolated load spectrum.

2. The method according to claim 1, characterized in that, The load distribution model includes a normal distribution model, where each load characteristic value includes the peak load of the powertrain mounting system under each preset operating condition; obtaining the load distribution model corresponding to each preset operating condition based on each load characteristic value includes: Obtain the pre-constructed coefficient of variation, and the correspondence between the peak load and the standard deviation and mean of the normal distribution model; the coefficient of variation is used to characterize the correlation between the standard deviation and the mean. For any of the preset working conditions, based on the load peak value, the coefficient of variation, and the correspondence between the load peak value and the standard deviation and mean value, the standard deviation and mean value of the normal distribution model corresponding to the preset working condition are obtained. Based on the standard deviation and mean of the normal distribution model corresponding to the preset working condition, construct the load distribution model for the preset working condition.

3. The method according to claim 2, characterized in that, The step of generating a load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition includes: For any of the preset working conditions, a Gaussian white noise sequence matching the predicted driving time is generated based on the predicted driving time corresponding to the preset working condition. Using the standard deviation and mean of the load distribution model corresponding to the preset working condition, the Gaussian white noise sequence is linearly transformed to obtain the original load sequence corresponding to the preset working condition. The original load sequence is input into a preset low-pass filter, and the load sequence corresponding to the preset operating condition is output through the low-pass filter.

4. The method according to claim 3, characterized in that, The original load sequence is composed of original load values ​​corresponding to multiple sampling points; the step of outputting the load sequence corresponding to the preset operating condition through the low-pass filter includes: Obtain the filter coefficients corresponding to the low-pass filter, and the original load value corresponding to the current sampling point; the current sampling point is any one of the multiple sampling points. When the current sampling point is the first sampling point among the plurality of sampling points, the original load value corresponding to the current sampling point is used as the filtered load value of the current sampling point; If the current sampling point is not the first sampling point, the filtered load value of the previous sampling point of the current sampling point is obtained, and the filtered load value of the previous sampling point and the original load value corresponding to the current sampling point are weighted using the filter coefficients to obtain the filtered load value of the current sampling point. The load sequence corresponding to the preset working condition is constructed using the filtered load values ​​of each current sampling point.

5. The method according to claim 1, characterized in that, The step of obtaining the predicted driving time of a vehicle equipped with the powertrain mounting system under each of the preset operating conditions includes: The mileage percentage of each preset working condition in the entire life cycle of the vehicle is obtained from the load report, and the driving speed of the vehicle in each preset working condition is obtained. Based on the mileage percentage of each preset working condition and the driving speed corresponding to each preset working condition, the driving time percentage corresponding to each preset working condition is obtained. Based on the predicted total driving time of the vehicle under each preset working condition and the percentage of driving time under each preset working condition, the predicted driving time corresponding to each preset working condition is obtained.

6. The method according to claim 1, characterized in that, The process of generating the extrapolated load spectrum associated with the powertrain mounting system based on the extrapolated load-time history includes: Rainflow counts are performed on the extrapolated load-time history, and complete load cycles contained in the extrapolated load-time history are extracted, along with the cycle amplitude and cycle mean of each complete load cycle. Obtain the amplitude range of each of the cyclic amplitudes and the mean range of each of the cyclic mean values, and divide multiple amplitude-mean intervals based on the amplitude range and mean range; Based on the cyclic amplitude and cyclic mean of each complete load cycle, the cyclic frequency matching each amplitude-mean interval is obtained, a rainflow matrix is ​​constructed according to the cyclic frequency matching each amplitude-mean interval, and the rainflow matrix is ​​used as the extrapolated load spectrum.

7. The method according to any one of claims 1 to 6, characterized in that, The robust optimization of the powertrain mounting system using the extrapolated load spectrum includes: Using the extrapolated load spectrum, the fatigue life assessment value of the powertrain mounting system under different combinations of design parameters is obtained; Based on the various combinations of design parameters and the fatigue life assessment values, a Kriging approximation model is constructed; the Kriging approximation model is used to characterize the correlation between the combinations of design parameters and the fatigue life assessment values. Using the Kriging approximation model, the first combination of design parameters corresponding to the powertrain mounting system is obtained when the fatigue life assessment value is optimal. If the first combination of design parameters satisfies the preset robustness conditions, the first combination of design parameters shall be used as the optimized combination of design parameters for the powertrain mounting system. If the first combination of design parameters does not meet the preset robustness condition, the optimal fatigue life assessment value is obtained by using the Kriging approximation model, and if the preset robustness condition is met, the second combination of design parameters corresponding to the powertrain mounting system is used as the optimized design parameter combination of the powertrain mounting system.

8. The method according to claim 7, characterized in that, The process of obtaining fatigue life assessment values ​​of the powertrain mounting system under different combinations of design parameters using the extrapolated load spectrum includes: For any combination of design parameters, a finite element model of the powertrain mounting system under the design parameter combination is constructed, and a unit load is applied to the finite element model to perform static analysis, and the static stress response results of the powertrain mounting system under the design parameter combination are output. Based on the static stress response results and the extrapolated load spectrum, the fatigue life assessment value of the powertrain mounting system under the design parameter combination is generated.

9. An optimization device for a powertrain mounting system, characterized in that, The device includes: The load information acquisition module is used to acquire the load report associated with the powertrain mounting system to be optimized, acquire the load characteristic value of the powertrain mounting system under each preset working condition from the load report, and acquire the predicted driving time of the vehicle equipped with the powertrain mounting system under each preset working condition. The load sequence generation module is used to obtain the load distribution model corresponding to each preset working condition based on the load feature value, and generate the load sequence corresponding to each preset working condition based on the load distribution model corresponding to each preset working condition and the predicted driving time corresponding to each preset working condition. The extrapolation load spectrum construction module is used to splice the load sequences corresponding to each of the preset working conditions, generate the extrapolation load-time history associated with the powertrain mounting system, and generate the extrapolation load spectrum associated with the powertrain mounting system based on the extrapolation load-time history. The robustness optimization module is used to perform robustness optimization processing on the powertrain mounting system using the extrapolated load spectrum.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images