Methods for monitoring the condition of mechanical systems, the system itself, and computer-readable storage media

Abstract

This application discloses a condition monitoring method for a mechanical system, the system itself, and a computer-readable storage medium. This condition monitoring method is used to monitor the health status of a mechanical system, and includes acquiring a... i The real-time operating load is calculated using a damage calculation model to obtain a. i The system accumulates damage in real time and calculates the real-time wear and tear life of the AI. Through the damage calculation model, it can obtain real-time damage conditions over multiple time periods, providing a universally applicable and highly reliable standard parameter for judging overall damage. Combined with the AI's preset weighting coefficients, the real-time comprehensive wear and tear life of the mechanical system is calculated, which is more conducive to judging the safety performance of the entire mechanical system. By analyzing the health status under various conditions, real-time judgment of the entire mechanical system can be achieved, enabling proactive contact with users for maintenance of specific components or the entire mechanical system, thus preventing safety accidents from occurring in advance.

Description

Technical Field

[0001] This application relates to the field of mechanical inspection, and in particular to a method for monitoring the condition of a mechanical system, a mechanical system including the method for monitoring the condition of the mechanical system, and a computer-readable storage medium. Background Technology

[0002] Compared to traditional vehicles, electric vehicle drive motors have a wider speed range, higher starting torque, higher power density, and higher efficiency. However, the increased load intensity, such as speed and torque, correspondingly exacerbates the risk of electric drive system failure. Furthermore, the lower operating costs of electric vehicles lead to increased daily mileage and more frequent usage, thus requiring electric drive systems with longer lifespans.

[0003] However, with the rapid development of automotive electrification and the increasing intelligence of electric drive systems, the current level of intelligence is too simplistic. It cannot apply precise control algorithms to damage monitoring of mechanical components in electric drive systems, and therefore cannot prevent component failures caused by the mechanical system reaching its lifespan limit. Summary of the Invention

[0004] This application provides a method for monitoring the condition of a mechanical system, a system for monitoring the condition of a mechanical system, and a computer-readable storage medium. This addresses the problems in existing technologies where real-time monitoring of vehicle mechanical systems is impossible, and where users cannot be given early warnings about system lifespan.

[0005] One aspect of this application provides a method for monitoring the condition of a mechanical system, used to monitor the health status of the mechanical system, wherein the mechanical system includes n different components, and the n different components are a i Where i is an integer between 1 and n, including: obtaining the a i The real-time operating load; according to the a i The real-time operating load is calculated using a damage calculation model to obtain a. i Real-time cumulative damage. According to a i Real-time cumulative damage and the a i The preset full-lifecycle damage target is used to calculate the a. i The real-time wear and tear lifetime. According to the aforementioned a i Real-time wear and tear lifetime and the a i The real-time comprehensive wear and tear life of the mechanical system is calculated using preset weighting coefficients. Based on the real-time comprehensive wear and tear life and the preset wear and tear range, the corresponding health status of the mechanical system is obtained.

[0006] Wherein, the a i Real-time cumulative damage and the a iThe preset full-lifecycle damage target is used to calculate the a. i The real-time wear and tear lifetime includes: via the formula: The a was calculated i The real-time wear and tear lifetime. Among them, D ai For the a i Real-time cumulative damage; D ti For the a i Preset full life cycle damage target; L ai For the a i The real-time wear and tear of the lifetime.

[0007] Wherein, according to the a i The real-time operating load is calculated using a damage calculation model to obtain the value of a. i Real-time cumulative damage, including: acquiring the a i The a in the real-time running load i The failure-dominant load. According to the aforementioned a i Based on the structural characteristics and failure modes, the dominant failure loads of the ai were selected for counting and statistics. i The preset SN curve and the a i The failure-dominant load was calculated, and the value of a under different load levels was obtained. i The damage coefficient is calculated based on the preset SN curve and Miner damage model, combined with the statistical results of the count of the failure-dominant load. i Real-time cumulative damage.

[0008] Wherein, the collection of a i The failure-dominant load includes: data acquired by sensors from the a i The real-time operating load. According to the aforementioned a i The structural characteristics and failure modes of the a were obtained. i Real-time data of the failure-dominant load.

[0009] Wherein, according to the a i Real-time wear and tear lifetime and the a i The real-time comprehensive wear life of the mechanical system is calculated using a preset weighting coefficient, including: via the formula: L m =k1L a1 +k2L a2 +…+k n L an The real-time comprehensive wear and tear life of the mechanical system is calculated; where k i Represents the a i Preset weighting coefficients; L ai Represents the a iReal-time wear life; n is the number of parts; L m This refers to the real-time comprehensive wear and tear life of the mechanical system.

[0010] The calculation steps for the preset weight coefficients include: obtaining the a values ​​of multiple samples. i The real-time wear and tear lifetime is calculated, and multiple values ​​of a are obtained. i The mean and standard deviation of the real-time wear and tear lifetime. This is determined using the formula: The a was calculated i The coefficient of variation; where σ represents the a i The standard deviation of the real-time wear lifetime; μ represents the a i The average real-time wear and tear lifetime; CV represents the value of a. i The coefficient of variation. Derived from the formula: The preset weighting coefficients are calculated.

[0011] Wherein, the a i The steps for obtaining the preset full-life-cycle damage target include: collecting multiple sample operation data of the mechanical system over a certain period of time, and selecting a preset sample from the multiple samples according to preset conditions. The steps also include obtaining a from the sample. i The real-time operating load. The damage calculation model is used to calculate the value of a in all samples. i The cumulative damage over a certain period of time is processed to obtain the a in all samples. i The target of damage. According to a i The preset scrapping parameters and the a i The ratio of the current parameters corresponding to the damage target, and the ratio of a in all samples. i The damage target is multiplied by the corresponding ratio to obtain the a. i The preset full life cycle damage target.

[0012] Among them, the a in all samples is calculated by the damage calculation model. i The real-time cumulative damage is processed to obtain the a in all samples. i The damage target includes: according to the a i The structural characteristics and failure modes of the components were collected, and the data of component a were analyzed. i The failure-dominant load. According to the aforementioned a i Based on the structural characteristics and failure modes, select the a i The corresponding preset load counting method is used for the a i The real-time operating load is counted and statistically analyzed, and the damage calculation model is used to calculate the value of a. i The maximum stress. Based on the preset SN curve and Miner's cumulative damage model, the value of a is calculated. iReal-time cumulative damage.

[0013] Wherein, the step of obtaining the a in the sample i The real-time operating load is calculated using a damage calculation model to obtain a in all samples. i The real-time cumulative damage calculation includes: collecting the dominant failure load of the ai based on the structural characteristics and failure modes of the ai components; selecting a preset load counting method corresponding to the ai based on the structural characteristics and failure modes of the ai to count and statistically analyze the real-time operating load of the ai; calculating the maximum stress of the ai based on the damage calculation model; and calculating the real-time cumulative damage of the ai based on the preset SN curve and the Miner cumulative damage model.

[0014] Specifically, based on the real-time comprehensive wear life and the preset wear range, the corresponding health status of the mechanical system is obtained, including: in response to the real-time comprehensive wear life being less than or equal to 50%, it is in a healthy state, and the real-time comprehensive wear life continues to be monitored in real time.

[0015] The condition monitoring method further includes a risk state response when the real-time comprehensive loss lifetime exceeds 50%. Based on the real-time comprehensive loss lifetime and a preset value range, the risk state is classified into different levels, and corresponding decision recommendations are provided according to different levels.

[0016] Specifically, a low-risk scenario is defined as follows: if the real-time comprehensive wear and tear lifespan is greater than 50% and less than or equal to 70%, regular maintenance of the mechanical system is required. A medium-risk scenario is defined as follows: if the real-time comprehensive wear and tear lifespan is greater than 70% and less than or equal to 90%, in-depth overhaul of the mechanical system is required. A high-risk scenario is defined as follows: if the real-time comprehensive wear and tear lifespan is greater than 90%, replacement of parts in the mechanical system is required.

[0017] One aspect of this application provides a condition monitoring system for a mechanical system, the monitoring system including a processor and a memory coupled to the processor, the memory storing program instructions for implementing the condition monitoring method as described in any of the preceding claims. The processor is configured to execute the program instructions in the memory to implement the condition monitoring method as described in any of the preceding claims.

[0018] Another aspect of this application provides a computer-readable storage medium that can be read by a processor and stores a program file capable of implementing the state monitoring method as described in any of the preceding claims, the program file being executable by a processor to implement the state monitoring method as described in any of the preceding claims.

[0019] Compared to existing technologies, the mechanical system condition monitoring method of this application is used to monitor the health status of a mechanical system, wherein the mechanical system includes n different components, and the n different components are a i where i is an integer between 1 and n, including obtaining a i Real-time operating load. Capable of recording the operating data of multiple components within a mechanical system over multiple time periods. Based on a i The real-time operating load is calculated using a damage calculation model to obtain a. i The real-time cumulative damage, after being processed by the damage calculation model, can yield real-time damage conditions over multiple time periods, providing a numerical basis for obtaining more reliable lifetime data. According to a... i Real-time cumulative damage and a i The preset full life cycle damage target is used to calculate a. i The real-time wear life was obtained, and the wear life obtained by judging the damage values ​​of internal components of the mechanical system over multiple time periods was obtained, providing a common and highly reliable standard parameter for judging overall damage. According to a i Real-time wear life and a i By using preset weighting coefficients, the real-time comprehensive wear life of the mechanical system is calculated. Calculating the overall wear life of the mechanical system by analyzing the wear life of multiple components provides a more accurate assessment of its safety performance. Based on the real-time comprehensive wear life and preset wear range, the corresponding health status of the mechanical system is obtained. By analyzing the health status under each condition, real-time assessment of the entire mechanical system is achieved, enabling proactive contact with users for maintenance of specific components or the entire mechanical system, thus preventing potential safety incidents. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating an embodiment of the mechanical system condition monitoring method provided in this application, wherein the path planning method includes steps S2 and S5;

[0021] Figure 2 yes Figure 1 A flowchart of an embodiment of step S2 described above includes step S21;

[0022] Figure 3 yes Figure 2 A flowchart illustrating an embodiment of step S21 described above;

[0023] Figure 4 yes Figure 1 A flowchart illustrating an embodiment of step S40 prior to step S4 in the mechanical system condition monitoring method described above;

[0024] Figure 5 yes Figure 1A flowchart of an embodiment of step S30 prior to step S3 described above, including steps S302 and S303;

[0025] Figure 6 yes Figure 5 A flowchart illustrating an embodiment of step S303 described above;

[0026] Figure 7 yes Figure 5 A flowchart illustrating an embodiment of step S302 described above;

[0027] Figure 8 yes Figure 1 A flowchart illustrating an embodiment of step S5 described above;

[0028] Figure 9 This application describes an embodiment of the distribution of average daily mileage across all users in a specific application scenario.

[0029] Figure 10 This is an embodiment of the present application showing the distribution of user-integrated wear and tear lifetime in an application scenario. Detailed Implementation

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] In the following description, specific details such as particular system architectures, interfaces, and technologies are presented for illustrative purposes rather than for limiting purposes, in order to provide a thorough understanding of this application.

[0032] In this paper, the terms "system," "unit," and "network" are often used interchangeably. The term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, "many" in this paper means two or more.

[0033] One aspect of this application provides a method for monitoring the state of a mechanical system, used to monitor the health state of the mechanical system, wherein the mechanical system includes n different components, and the n different components are a i Let i be an integer between 1 and n, and include the following steps:

[0034] Step S1: Obtain a iThe system can record real-time operating loads. Through the cloud-based system, the operating loads of various mechanical components over multiple time periods can be recorded in real time, including torque, speed, temperature, current, voltage, etc. Acquiring real-time operating loads facilitates analysis of the mechanical system's operation during subsequent maintenance and enables monitoring of the system's internal status at various stages. This avoids the limitations of existing technologies that rely solely on mechanical system faults for system-wide troubleshooting, leading to inadequate or delayed maintenance, thus improving reliability and safety.

[0035] Step S2: According to a i The real-time operating load is calculated using a damage calculation model to obtain a. i The real-time cumulative damage, after being processed by the damage calculation model, can be obtained for each component over multiple time periods, i.e., a. i The real-time cumulative damage data fully utilizes the cloud's data acquisition and storage capabilities, providing a numerical foundation for obtaining more reliable lifespan data in the future.

[0036] Step S3: According to a i Real-time cumulative damage and a i The preset full life cycle damage target is used to calculate a. i The system calculates and determines the real-time wear-out lifespan of components by obtaining the cumulative damage of corresponding components over a period of time, and then combining this with existing AI-preset full-lifecycle damage targets. This process transforms the damage of different components into a pre-set reference standard and converts the real-time wear-out lifespan into a unified variable through a specific conversion method, providing a universally applicable and highly reliable standard parameter for assessing overall health status.

[0037] In some application scenarios, by obtaining the real-time cumulative damage of multiple corresponding components in the cloud mechanical system of the sample user, the real-time loss life can be calculated based on the preset full life cycle damage target corresponding to multiple components.

[0038] Step S4: According to a i Real-time wear life and a i The preset weighting coefficients are used to calculate the real-time comprehensive wear life of the mechanical system. By comprehensively calculating the wear life of the entire mechanical system through the actual wear life of multiple components, the mechanical system in use can be monitored, which is more conducive to judging the safety performance of the entire mechanical system and improving the safety of users during use.

[0039] Step S5: Based on the real-time comprehensive wear life and the preset wear range, obtain the corresponding health status of the mechanical system. By analyzing the health status under each condition, real-time assessment of the entire mechanical system can be achieved. Furthermore, the real-time data from the above steps can be stored in the cloud, facilitating direct troubleshooting and handling of faulty structures within the entire mechanical system. Simultaneously, it enables proactive contact with users to perform maintenance on specific components or the entire mechanical system, preventing safety accidents from occurring in advance.

[0040] In order to enable data to be recorded in the cloud in real time and to calculate component a using a preset formula i The implementation wear and tear life. In some specific embodiments, step S3 above includes:

[0041] Through the formula: Calculate a i Real-time wear and tear lifespan.

[0042] Among them, D ai For a i Real-time cumulative damage; D ti For a i Preset full life cycle damage target; L ai For a i The lifetime of real-time wear and tear. This is calculated by a. i Real-time cumulative damage accounts for the entire a i The proportion of the preset full-lifecycle damage target can be used to obtain a. i The real-time wear and tear life is calculated, and the wear of each component is quantified using a specific standard parameter.

[0043] In order to better leverage the combination of cloud platforms and existing technologies, in acquiring a i The real-time cumulative damage is calculated using different data carriers for different components, and then different calculation methods are applied. Please refer to [the relevant documentation]. Figure 2 , Figure 2 yes Figure 1 A flowchart illustrating an embodiment of step S2 described above. Step S2 specifically includes:

[0044] Step S21: Collect a i a in the real-time running load i The dominant failure load. In a mechanical system, due to the various types of components, each component has different critical points and maximum stress points; therefore, the calculation methods for the maximum stress of different components also differ. After obtaining the real-time operating load of a component, the cloud system automatically matches and obtains the dominant failure load of that component based on its characteristics and selects the corresponding damage calculation model in the cloud for calculation.

[0045] In specific implementation scenarios, when the components acquired in a mechanical system are shafts, the main failure mode of these shafts is fatigue failure, which can be categorized into tensile / compressive fatigue, bending fatigue, and torsional fatigue. These tend to initiate at stress concentration points, surfaces, or subsurfaces, respectively. In mechanical systems, shafts primarily bear torque, and their failure mode is torsional fatigue failure. The damage depends on the varying torque loading history under different load conditions. Given that motor shafts under actual operating conditions are mainly subjected to torsional fatigue due to torque, tensile / compressive fatigue and bending fatigue have negligible impacts on shaft damage. Therefore, the dominant load for shaft failure is shear stress.

[0046] Step S22: According to a i Based on the structural characteristics and failure modes, select a i The corresponding preset load counting method will a i The failure-dominant load was calculated, and a was obtained. i The maximum stress. Based on the above determination that the component is a shaft component and the calculation of shear stress, it needs to be simplified that the motor shaft is a shaft with a constant cross-section and a diameter of D. Therefore, the calculation formula is:

[0047]

[0048] Given the actual torque T of the reducer input shaft, the torque on the motor shaft and each bearing of the reducer can be calculated based on the transmission ratio, as shown in Table 1.

[0049] Table 1 Torque Decomposition Table for Shaft Components

[0050]

[0051] Among them, z x This represents the number of teeth on the x-th helical gear.

[0052] Considering the influence of the mean shear stress on shaft damage, the relationship between the mean shear stress and the amplitude of shear stress is obtained according to the Goodman linear formula as follows, and the amplitude can be corrected by the mean stress correction formula.

[0053]

[0054] Among them, S R The equivalent zero-mean shear stress amplitude (MPa); S a The shear stress amplitude (MPa)

[0055] It yielded 663 MPa.

[0056] Step S23: According to a i The preset SN curve and a i The failure-dominant load was calculated, and a was obtained under different load levels. iThe damage coefficient is the relationship between the maximum stress and fatigue life of a component, and the damage coefficient varies under different load levels.

[0057] In this implementation scenario, the slope and intercept of the SN curve are set using existing empirical parameters, thereby avoiding the drawback that a large number of experiments are required to obtain the SN curve in actual applications due to different component materials.

[0058] Step S24: Based on the preset SN curve and Miner damage accumulation model, the real-time cumulative damage of ai is calculated. Therefore, for the failure mode of shaft parts, the dominant load for failure is torque, and the fatigue life N under the i-th torque range can be calculated. f,i After counting the load rainflow cycles, the i-th stage torque range (Range) S is obtained. R,i The number of loops n R,i .

[0059] According to Miner's linear cumulative damage criterion, the damage to the shaft component caused by each running segment can be calculated as D. s The calculation formula is as follows:

[0060]

[0061] In order to better integrate cloud data with existing technologies, in acquiring a i The real-time cumulative damage is calculated using different data carriers for different components, and then different calculation methods are applied. Please refer to [the relevant documentation]. Figure 3 , Figure 3 yes Figure 2 A flowchart illustrating an embodiment of step S21 described above. Step S21 specifically includes:

[0062] Step S211: Collect the real-time operating load of the AI ​​through sensors. In some implementation scenarios, when collecting real-time data of the entire mechanical system, it is generally done by corresponding sensors installed within the system. The real-time operating load includes, but is not limited to, torque, rotational speed, vehicle speed, temperature, current, voltage, etc.

[0063] Step S212: Based on the structural characteristics and failure mode of the ai, obtain real-time data of the failure-dominant load of the ai, and analyze and process it to obtain its maximum stress. In some implementation scenarios, components in a mechanical system can be divided into shaft parts and rotating parts. Among them, shaft parts, as described in the above embodiments, are mainly subjected to torsional fatigue due to torque, and their maximum stress point should be the stress concentration point on the shaft; the load on rotating parts is related to the torque fluctuation of the input shaft on the one hand, and the frequency of load application is related to the rotational speed of the input shaft on the other hand. Therefore, the maximum stress point of rotating parts such as gears should be the contact point of the gear pair and the tooth root fillet. Specifically, for components in a mechanical system, their real-time operating loads are mostly torque and rotational speed, and the stress that leads to their failure is contact stress. For shaft parts, their stress is as described in Formula 1. For rotating parts, taking gears as an example, the number of gear rotations at a given torque level can be obtained according to Formula 4, then at a given torque level k i The number of revolutions of the lower gear is r i :

[0064]

[0065] Where, rotational speed, i = 1, 2, ..., N k To cover all torque levels k i j = 1, 2, ..., N n To cover all speed levels n j .

[0066] The contact stress of the gear is calculated based on the gear strength calculation formula of ISO 6336 standard, combined with torque, gear parameters and other conditions.

[0067]

[0068] Among them: Z B : Coefficient of engagement; Z H : Nodal region coefficient; Z E : Elastic modulus; Z ε Overlap coefficient; Z β : Helix angle factor; F T Nominal tangential force, the force acting on the pitch circle (N); u: gear ratio of the large gear / small gear; b: tooth width (mm); d: pitch circle diameter (mm); K A Use load factor; K V Dynamic load factor; K Hβ Tooth load distribution factor; K Hα : Inter-tooth load distribution coefficient.

[0069] In some implementation scenarios, to assess the health of a mechanical system holistically, it's necessary to combine the various components and sum them according to their predetermined proportions to arrive at a comprehensive value. This value is obtained in the cloud. i Real-time wear life and a i The preset weighting coefficients are determined by the formula:

[0070] L m =k1L a1 +k2L a2 +…+k n L an Formula Six

[0071] The real-time comprehensive wear and tear life of the mechanical system is calculated. Where, k i Represents a i Preset weighting coefficients; L ai Represents a i Real-time wear life; n is the number of parts; L m This refers to the real-time comprehensive wear and tear life of the mechanical system. Through L... m The calculation can yield the real-time comprehensive wear and tear life of the entire mechanical system, facilitating the evaluation of the entire mechanical system. Furthermore, the overall mechanical system values ​​are more in line with user requirements and easier for users to understand.

[0072] Regarding the above calculation L m The value of k, where k n The value is not random, but requires certain standards. This is to better calculate the real-time comprehensive wear life L of the mechanical system. m For details, please refer to Figure 4 , Figure 4 yes Figure 1 A flowchart illustrating an embodiment of step S40 prior to step S4 in the state monitoring method for the mechanical system described above. Step S40 specifically includes:

[0073] Step S401: Obtain the a of multiple samples i The real-time wear and tear lifetime was calculated, and multiple a values ​​were obtained. i The mean and standard deviation of the real-time wear and tear lifetime.

[0074] To avoid abnormal data in the operational data collected by the mechanical system or due to other erroneous operations, when acquiring sample data, the same type of mechanical system should be selected, and the real-time wear life of the same component from multiple sample users should be obtained within the same time period. The mean and standard deviation of these real-time wear lives should be calculated, and some abnormal data should be excluded, thereby reducing the overall calculation error.

[0075] Step S402: Using the formula: Calculate a iThe coefficient of variation; where σ represents a i The standard deviation of the real-time wear lifetime. Where μ represents a i The average real-time wear lifetime; CV represents a i The coefficient of variation.

[0076] Step S403: Using the formula: The preset weighting coefficients are calculated. In the evaluation index system, the greater the difference in index values, the more difficult it is to achieve, and therefore, higher weights are assigned to those values. Thus, in this method, components with greater lifespan deterioration are more prone to failure and reach their lifespan limit first, and are therefore assigned higher weights.

[0077] This method compares the lifespan and damage level of the mechanical system to be monitored and its components with a certain sample number of mechanical systems and their components. Therefore, it requires acquiring data from a sufficient number of mechanical systems and their components as a standard, i.e., setting a pre-defined full-life-cycle damage target. Please refer to [reference needed]. Figure 5 , Figure 5 yes Figure 1 A flowchart illustrating an embodiment of step S30 prior to step S3 described above. Step S30: Obtain a i The preset full life cycle damage targets specifically include:

[0078] Step S301: Collect multiple sample operation data of the mechanical system within a certain period of time, and select a preset sample from the multiple samples according to preset conditions.

[0079] In one application scenario, when acquiring data from different components, multiple user samples can be obtained, each containing real-time operating loads within the same time period, such as torque, rotational speed, vehicle speed, temperature, current, and voltage. Based on preset conditions and the needs of the components and other factors, load data that can be used as preset samples can be selected from these loads.

[0080] Understandably, each sample's mechanical system also includes n different parts, where the n different parts are a. i , where i is an integer between 1 and n.

[0081] In some application scenarios, please refer to Figure 9 , Figure 9This application describes the distribution of average daily mileage for all users in a specific application scenario, according to one embodiment of this application. Since users with different average daily mileages experience significantly different driving conditions, for example, users with an average daily mileage of over 100km are mostly long-distance commuters, spending most of their time on highways and overpasses, while users with an average daily mileage of only a dozen kilometers are short-distance commuters, primarily driving on urban roads. Therefore, considering the varying damage to components caused by different driving conditions, the selected user sample should comprehensively cover users in different driving scenarios to ensure that the determined user lifecycle damage target is more realistic.

[0082] Step S302: Obtain a from the sample i The real-time operating load was calculated. After basic preprocessing of the selected user data, removing outliers and missing values, the cumulative damage to each component of each user sample was calculated. In some application scenarios, these components may include motor shafts, reducer gears, and bearings, and the cumulative damage to each component is shown in Table 2.

[0083] Table 2 Cumulative Damage to Different Parts by Users

[0084]

[0085] In some application scenarios, when calculating the real-time operational data of each component of a sample user, it is also necessary to calculate the real-time cumulative damage of each component using a damage calculation model. Please refer to [the relevant documentation]. Figure 7 , Figure 7 yes Figure 5 A flowchart illustrating an embodiment of step S302 described above, the specific steps of which include:

[0086] Step S3021: According to a i The structural characteristics and failure modes of the components were analyzed, and data collection was performed on a. i The failure-dominant load data.

[0087] Step S3022: According to a i Based on the structural characteristics and failure modes of the components, the corresponding preset load counting method is selected to count a. i The loads implemented during operation are counted and statistically analyzed, and the maximum stress of ai is calculated based on the damage calculation model.

[0088] Step S3023: Based on the preset SN curve and Miner cumulative damage model, calculate a. i Real-time cumulative damage.

[0089] Step S303: Calculate a in the sample using the damage calculation model. i The real-time cumulative damage is calculated and processed to obtain a from all samples. iThe damage target. To obtain data that better meets the actual needs of users, further screening of the sample data is required. Please refer to... Figure 6 , Figure 6 yes Figure 5 A flowchart illustrating an embodiment of step S303 described above, the specific steps of which include:

[0090] Step S3031: Based on a in all samples i Real-time cumulative damage, statistics of all a i The real-time cumulative damage distribution is obtained by arranging the acquired real-time cumulative damage by size and statistically analyzing the real-time cumulative damage distribution of each component.

[0091] In the above application scenario, when processing samples of mechanical systems, in a specific embodiment, 270 sample users can be selected to obtain the real-time cumulative damage of each component in the mechanical system of the 270 sample users over a period of time. Taking the motor shaft as an example, the cumulative damage of the motor shaft is obtained for three consecutive months. The obtained cumulative damage data is arranged according to size to obtain the real-time cumulative damage distribution of the motor shaft.

[0092] Step S3032: Based on all a i The damage distribution, selecting each a i The preset distribution ratio is used, and the real-time cumulative damage corresponding to the selected preset distribution ratio is taken as each a. i The damage target is determined. In the above application scenario, the real-time cumulative damage distribution of the motor shaft is obtained according to step S3031. The motor shaft damage of all sample users follows a 3-parameter logarithmic distribution. The damage value of 1.397E-7 at the 95th percentile is selected as the damage target of the sample users.

[0093] Step S304: According to a i The preset scrapping parameters and a i The ratio of the current parameters corresponding to the damage target, and a in all samples i The damage target is multiplied by the corresponding ratio to obtain a. i The preset full life cycle damage target is determined. In the above application scenario, taking the motor shaft as an example, based on the mileage distribution of sample users, the mileage at the 95th percentile is determined to be 13284.673km. The total life cycle mileage of the vehicle is set at 300,000 kilometers, resulting in an extrapolation factor of 22.58. The full life cycle damage target of the motor shaft is obtained by multiplying the sample user damage target by the extrapolation factor.

[0094] Table 3 shows the user-defined life-cycle damage targets for some parts obtained using this method.

[0095] Table 3 Damage Targets Throughout User's Life Cycle

[0096]

[0097] Furthermore, in some application scenarios, to provide users with an intuitive reference value for the entire mechanical system, please refer to... Figure 8 , Figure 8 yes Figure 1 A flowchart illustrating an embodiment of step S5 described above. Step S5 specifically includes:

[0098] Step S51: Obtain the corresponding health status of the mechanical system and check whether the real-time comprehensive wear and tear life is less than or equal to 50%.

[0099] If the real-time comprehensive wear and tear life is less than or equal to 50%, the mechanical system is considered healthy, and real-time monitoring of the comprehensive wear and tear life continues. An early warning can be issued when the real-time comprehensive wear and tear life of the mechanical system exceeds 50%.

[0100] Step S52: In response to a real-time comprehensive wear life exceeding 50%, the mechanical system is in a risky state. When the mechanical system is in a risky state, the user or inspector needs to pay closer attention to the entire mechanical system and further determine whether repair or replacement of parts is necessary to ensure user safety.

[0101] Step S53: Based on the real-time comprehensive loss life and the preset value range, determine whether the real-time comprehensive loss life is greater than 90%, classify the risk status, and when the mechanical system is in a risk status, give corresponding decision suggestions according to different levels.

[0102] If the real-time comprehensive wear and tear life is greater than 50% and less than or equal to 70%, it is considered low risk and requires in-depth overhaul and regular maintenance of the mechanical system.

[0103] A response indicating that the real-time comprehensive wear and tear life is greater than 90% is considered high risk and requires replacement of parts in the mechanical system.

[0104] Please refer to the reference. Figure 10 , Figure 10 This application presents an embodiment of the user's overall wear and tear life distribution in an application scenario. The corresponding health status and strategy recommendations for the entire mechanical system are shown in Table 4 below. In the above application scenario, the overall wear and tear life effectively reflects the user's driving habits. Even with the same mileage, the overall wear and tear life can vary significantly, and it increases as the user's mileage increases.

[0105] Table 4 Health Status Evaluation of Electric Drive Systems

[0106] Comprehensive wear and tear life Health status assessment After-sales strategy <![CDATA[L m ≤50%]]> Risk-free Continuous monitoring <![CDATA[50%<L m ≤70%]]> Low risk Regular maintenance <![CDATA[70%<L m ≤90%]]> Medium risk In-depth overhaul <![CDATA[90%<L m ]]> High risk Replacement parts

[0107] Another aspect of this application proposes a condition monitoring system for a mechanical system. The monitoring system includes a processor and a memory coupled to the processor. The memory stores program instructions for implementing the condition monitoring method as described in any of the above claims. The processor executes the program instructions in the memory to implement the condition monitoring method as described in any of the above claims.

[0108] Another aspect of this application proposes a computer-readable storage medium that can be read by a processor and stores a program file capable of implementing the state monitoring method as described above. The program file can be executed by the processor to implement the state monitoring method as described above.

[0109] This application discloses a method for monitoring the state of a mechanical system, used to monitor the health status of the mechanical system, wherein the mechanical system comprises n distinct components, and the n distinct components are a. i where i is an integer between 1 and n, including obtaining a i Real-time operational data. It can record the operational data of components over multiple time periods. According to a... i The real-time operating data is used to calculate a using a damage calculation model. i The real-time cumulative damage, after being processed by the damage calculation model, can yield real-time damage conditions over multiple time periods, providing a numerical basis for obtaining more reliable lifetime data. According to a... i Real-time cumulative damage and a i The preset full life cycle damage target is used to calculate a. i The real-time wear life was obtained, and the wear life obtained by judging the damage values ​​of internal components of the mechanical system over multiple time periods was obtained, providing a common and highly reliable standard parameter for judging overall damage. According to a i Real-time wear life and a i By using preset weighting coefficients, the real-time comprehensive wear life of the mechanical system is calculated. Calculating the overall wear life of the mechanical system by analyzing the wear life of multiple components provides a more accurate assessment of its safety performance. Based on the real-time comprehensive wear life and preset wear range, the corresponding health status of the mechanical system is obtained. By analyzing the health status under each condition, real-time assessment of the entire mechanical system is achieved, enabling proactive contact with users for maintenance of specific components or the entire mechanical system, thus preventing potential safety incidents.

[0110] In the description of this application, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0111] The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of condition monitoring of a mechanical system for monitoring a health condition of the mechanical system, wherein, The mechanical system comprises n different components, wherein the n different components are a i , i is an integer between 1 and n, characterized in that it comprises: Acquisition a i Real-time running load; According to the a i real-time running load, the real-time cumulative damage of the a i is calculated by a damage calculation model. According to the a i real-time cumulative damage and the a i preset life cycle damage target of the a i real-time loss life is calculated. According to the a i Real-time loss life and the a i Preset weight coefficient, the real-time comprehensive loss life of the mechanical system is calculated. According to the real-time comprehensive loss life and the preset loss range, a corresponding health state of the mechanical system is obtained.

2. The condition monitoring method of claim 1, wherein, The a i Real-time cumulative damage and the a i Pre-set life cycle damage target of the a i Real-time loss life, comprising: The real-time loss life of the a is calculated by the formula: i ; wherein, is the real-time cumulative damage of the a i ; is the preset total life cycle damage target of the a i ; is the real-time loss life of the a i .

3. The condition monitoring method of claim 1, wherein, The real-time running load of the a i The real-time cumulative damage of the a i The real-time cumulative damage of the a collecting the a i a i failure dominant load in real-time running load; According to the structural characteristics and failure mode of the a i , the a i corresponding preset load counting method is selected to count the failure dominant load of the a i . According to the a i preset S-N curve and the a i failure dominant load of the a i damage coefficient of the a According to the preset S-N curve and the Miner cumulative damage model, the real-time cumulative damage of the a i is calculated in combination with the counting statistical result of the failure dominant load.

4. The condition monitoring method of claim 3, wherein, said collecting the a i failure dominant load; Comprise: acquire the real-time running load of the a i through the sensor According to the a i Structural features and failure modes of the ai, real-time data of the failure dominant load of the ai is obtained.

5. The condition monitoring method according to any one of claims 1-4, characterized in that, The real-time loss life according to the a i The preset weight coefficient of the a i The real-time comprehensive loss life of the mechanical system is calculated by the following steps: Through the formula: The real-time comprehensive wear and tear life of the mechanical system is calculated; wherein, Represents the a i The preset weighting coefficients; Represents the a i The real-time wear life; n is the number of parts; This refers to the real-time comprehensive wear and tear life of the mechanical system.

6. The condition monitoring method of claim 5, wherein, The calculation step of the preset weight coefficient comprises: acquire the a of multiple samples i the mean and standard deviation of the real-time loss life of the a i of multiple samples Through the formula: The a was calculated i The coefficient of variation; where Represents the a i The standard deviation of real-time wear-out lifetime; Represents the a i The average real-time wear and tear lifetime; CV represents the value of a. i coefficient of variation; The preset weight coefficient of a is calculated by the formula: i ​​ 7. The condition monitoring method according to any one of claims 1-4, characterized by, The a i The obtaining step of the preset total life cycle damage target comprises: Collecting a plurality of sample running data of the mechanical system within a certain time, and selecting a preset sample from the plurality of samples according to a preset condition; acquiring a i real-time running load of a The cumulative damage of the a i in the certain time of all samples is calculated by the damage calculation model and processed to obtain the damage target of the a i in all samples; According to the a i The ratio of the preset scrap parameters of the a i The current parameters corresponding to the damage targets of the a i Multiply the damage targets of the a i The preset total life cycle damage targets of the a 8. The condition monitoring method of claim 7, wherein, The cumulative damage of the a within the certain time of all samples is calculated by the damage calculation model, and the damage target of the a of all samples is obtained by processing. i The cumulative damage of the a within the certain time of all samples is calculated by the damage calculation model, and the damage target of the a of all samples is obtained by processing. i The cumulative damage of the a within the certain time of all samples is calculated by the damage calculation model, and the damage target of the a of all samples is obtained by According to all samples, the real-time cumulative damage of a i The distribution of the real-time cumulative damage of a i The distribution of the real-time cumulative damage of a According to all the a i real-time cumulative damage distribution, select a preset distribution ratio of each a i , and take the real-time cumulative damage corresponding to the selected preset distribution ratio as the damage target of each a i .

9. The condition monitoring method of claim 7, wherein, The acquisition of the a i Real-time running load of the sample, the cumulative damage of the a i of all samples within a certain time is calculated by damage calculation model, comprising: According to a i Structural features and failure modes of components, collect a i Failure dominant loads; According to the structure characteristics and failure mode of the a i , select a i corresponding preset load counting method to count the real-time running load of the a i , and calculate the maximum stress of the a i according to the damage calculation model. According to the preset S-N curve and the Miner cumulative damage model, the real-time cumulative damage of the a i is calculated.

10. The condition monitoring method according to any one of claims 1-4, characterized by, According to the real-time comprehensive loss life and the preset loss range, a corresponding health state of the mechanical system is obtained, comprising: In response to the real-time comprehensive loss life being less than or equal to 50%, the health state is obtained, and the real-time monitoring of the real-time comprehensive loss life is continued.

11. The condition monitoring method of claim 10, wherein, Also includes in response to the real-time comprehensive loss life being greater than 50%, the risk state is obtained; According to the real-time comprehensive loss life and the preset value range, the risk state is classified, and corresponding decision suggestions are given according to different grades.

12. The state monitoring method of claim 11, wherein, In response to the real-time comprehensive loss life being greater than 50% and less than or equal to 70%, the low risk state is obtained, and the mechanical system needs to be maintained regularly; In response to the real-time comprehensive loss life being greater than 70% and less than or equal to 90%, the medium risk state is obtained, and the mechanical system needs to be overhauled deeply; In response to the real-time comprehensive loss life being greater than 90%, the high risk state is obtained, and the mechanical system needs to be replaced.

13. A condition monitoring system of a mechanical system, characterized by The monitoring system comprises a processor and a memory coupled to the processor, The memory stores program instructions for implementing the state monitoring method of any one of claims 1-12; The processor is configured to execute the program instructions in the memory to implement the state monitoring method of any one of claims 1-12.

14. A computer-readable storage medium, characterized in that, The program file can be read by the processor and can implement the state monitoring method of any one of claims 1-12, and the program file can be executed by the processor to implement the state monitoring method of any one of claims 1-12.

Images