Dynamic pressure oil film bearing thermal-mechanical performance prediction method and system based on artificial intelligence

Abstract

The invention discloses a dynamic pressure oil film bearing thermal-mechanical performance prediction method and system based on artificial intelligence, and the method is used for predicting the performance parameters of a dynamic pressure oil film bearing, and comprises the steps: obtaining the structure parameter data and working condition parameter data of a target dynamic pressure oil film bearing; and inputting the structure parameter data and the working condition parameter data of the target dynamic pressure oil film bearing into the trained artificial intelligence model to obtain the performance parameters of the target dynamic pressure oil film bearing. Through the method and the system, the traditional numerical calculation method based on an empirical formula or a finite element / finite difference and the like can be replaced, the prediction efficiency and precision are remarkably improved, and historical data and simulation data can be fully utilized.

Description

Technical Field

[0001] The present invention relates to the field of fluid bearing / intelligent algorithm technology, and more specifically to a method and system for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence. Background Art

[0002] Oil film bearing is a radial sliding bearing that uses lubricating oil as the lubricating medium. Among them, tilting pad hydrodynamic oil film bearings are widely used in high-speed and heavy-load industrial machinery such as turbomachinery, air compressors, electric spindles, and pumps due to their high load-bearing capacity and wide speed range. When the rotor rotates at high speed, a wedge-shaped oil film will be generated between the rotor and the bearing; under the dynamic pressure support of the oil film, the rotor is suspended in the bearing. However, the high-speed, high-load, and high-viscosity application environment characteristics are bound to cause varying degrees of fatigue loss to the bearing, which in turn causes the instability of the rotor system and even causes the entire rotating machinery equipment to shut down and scrap. Generally, the stability of the rotor-bearing system is related to performance parameters such as the load capacity of the bearing, the stiffness and damping of the bearing, the minimum oil film thickness, the temperature field distribution during the operation of the rotor-bearing, and the viscosity of the lubricating oil. There are many parameters that affect the above characteristics, such as oil supply temperature, bearing clearance, bearing aspect ratio, lubrication method, bearing pivot offset coefficient, and bearing preload.

[0003] The traditional numerical calculation method based on empirical formula or finite element / finite difference is to predict Dynamic oil film The main way to predict the mechanical properties and temperature field of bearings. Traditional numerical calculation methods have high accuracy in predicting the mechanical and thermodynamic properties of tilting pad bearings, but solving finite element models requires a lot of computing time, which greatly increases the computing cost. At the same time, some models are only applicable to a single configuration of bearing types, and the applicability of the model is low. Artificial intelligence is increasingly used in computational mechanics, combining traditional finite element mechanism models with machine learning to nonlinearly reduce the order of traditional models for fast and effective simulation. Although some studies have attempted to predict bearing characteristics through AI, the existing research still has the following shortcomings and cannot meet industrial needs: (1) Most of the existing bearing performance sample data are based on simplified bearing physical models, such as ignoring viscosity-temperature characteristics and bearing thermal elastic deformation, resulting in insufficient accuracy of bearing performance sample data and the failure of AI models to conduct accuracy verification; (2) Existing samples cover fewer bearing parameters, especially tilting pad bearings, which are the most widely used in the industry. They fail to fully cover bearing structure, operating conditions, lubricant viscosity, oil supply methods, etc. The predicted bearing characteristics are also very limited, and most of them are based on pressure distribution and load-bearing capacity prediction. The relatively small number of input and output nodes makes the AI ​​model too simple; (3) The relevant AI predictions only stay at the bearing performance prediction under fixed parameters and fail to carry out AI-based bearing structure parameter optimization for real rotor-bearing systems.

[0004] Therefore, there is an urgent need for a fast, accurate and effective reduced-order model to replace complex traditional numerical calculation methods based on empirical formulas or finite element / finite difference methods, so as to improve the efficiency and accuracy of bearing performance parameter prediction and solve the above technical problems. Summary of the invention

[0005] In view of this, the present invention provides a method and system for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence, which at least solves some of the technical problems in the background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] On the one hand, the present invention discloses a method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence, comprising the following steps:

[0008] Acquire structural parameter data and operating condition parameter data of a target hydrodynamic oil film bearing;

[0009] The structural parameter data and operating condition parameter data of the target hydrodynamic oil film bearing are input into the trained artificial intelligence model to obtain the performance parameters of the target hydrodynamic oil film bearing.

[0010] Furthermore, the structural parameters of the target hydrodynamic oil film bearing include: bearing wattage, clearance ratio, aspect ratio, preload and fulcrum type, etc.;

[0011] The operating parameters of the target dynamic pressure shaft seat include: rotation speed, load, lubrication method, load type, lubricating oil type, oil supply temperature and oil supply pressure.

[0012] Further, the performance parameters of the target hydrodynamic oil film bearing include mechanical properties and thermodynamic properties;

[0013] The thermodynamic characteristics include: maximum temperature, average temperature and power loss of the bearing;

[0014] The mechanical properties include: bearing stiffness, damping and minimum bearing oil film thickness.

[0015] Furthermore, the artificial intelligence model includes a deep learning neural network reduced-order model.

[0016] Furthermore, the artificial intelligence model is trained by the following steps:

[0017] Acquire a multi-source data set, wherein the multi-source data set includes hydrodynamic oil film bearing experimental test data, finite element analysis software simulation data, and bearing performance expert knowledge base data;

[0018] Acquire structural parameter data and operating parameter data of the hydrodynamic oil film bearing from the multi-source data set, and use them as data sets for training the deep learning neural network reduced-order model;

[0019] The data set is divided into a training set and a test set in a ratio of 4:1, and the training set is input into the deep learning neural network reduced-order model for training to obtain a trained deep learning neural network reduced-order model, and the deep learning neural network reduced-order model outputs the performance parameters of the hydrodynamic oil film bearing.

[0020] On the other hand, the present invention also discloses a thermal-mechanical performance prediction system of a hydrodynamic oil film bearing based on artificial intelligence, comprising:

[0021] A data input module, used for inputting structural parameter data and operating condition parameter data of a target dynamic pressure oil film bearing;

[0022] The performance prediction output module uses the trained artificial intelligence model and outputs the performance parameters of the target hydrodynamic oil film bearing based on the input structural parameter data and operating condition parameter data.

[0023] Preferably, in the data input module, the input structural parameters of the target hydrodynamic oil film bearing include: bearing wattage, clearance ratio, aspect ratio, preload and fulcrum type, etc.;

[0024] The operating parameters of the target dynamic pressure shaft seat include: rotational speed, load, lubrication method, load type, lubricating oil type, oil supply temperature and oil supply pressure.

[0025] Preferably, in the performance prediction output module, the output performance parameters of the target hydrodynamic oil film bearing include mechanical properties and thermodynamic properties;

[0026] The thermodynamic properties include: the maximum temperature, average temperature and power loss of the bearing;

[0027] The mechanical properties include: bearing stiffness, damping and minimum bearing oil film thickness.

[0028] Preferably, in the performance prediction output module, the artificial intelligence model includes a deep learning neural network reduced-order model.

[0029] Preferably, in the performance prediction output module, the artificial intelligence model is trained by the following steps:

[0030] Acquire a multi-source data set, wherein the multi-source data set includes hydrodynamic oil film bearing experimental test data, finite element analysis software simulation data, and bearing performance expert knowledge base data;

[0031] Acquire structural parameter data and operating condition parameter data of the hydrodynamic oil film bearing from the multi-source data set, and use them as data sets for training the deep learning neural network reduced-order model;

[0032] The data set is divided into a training set and a test set in a ratio of 4:1, and the training set is input into the deep learning neural network reduced-order model for training to obtain a trained deep learning neural network reduced-order model, and the deep learning neural network reduced-order model outputs the performance parameters of the hydrodynamic oil film bearing.

[0033] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence, which has the following beneficial effects:

[0034] The present invention adopts an artificial intelligence model to replace traditional numerical calculation methods based on empirical formulas or finite element / finite difference methods to realize bearing performance prediction. While ensuring the prediction accuracy, it effectively improves the calculation efficiency and reduces the calculation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0036] Figure 1 The present invention provides an overall schematic diagram of the method process. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] The embodiment of the present invention discloses a method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence, such as Figure 1 As shown, it includes: generation of multi-source data sets, construction of bearing structure-operating condition-performance mapping database, training of artificial intelligence models, and module integration and application of new models;

[0039] The sources of the multi-source data set include: experimental test data, hydrodynamic oil film bearing finite element analysis software and bearing performance expert knowledge base, the data set input includes structural parameters and operating condition parameters of the tilting pad hydrodynamic oil film bearing, and the output includes mechanical properties, thermal properties and other parameters of the bearing;

[0040] Furthermore, the structural parameters of the tilting pad dynamic oil film bearing include but are not limited to: bearing wattage, clearance ratio, aspect ratio, preload, fulcrum type, etc.; the operating parameters include but are not limited to: speed, load, lubrication method, load type, lubricating oil type, oil supply temperature, oil supply pressure, etc.;

[0041] Furthermore, the bearing mechanical properties include but are not limited to: bearing stiffness, damping and minimum bearing oil film thickness; the bearing thermal properties include but are not limited to: bearing maximum temperature, average temperature and power loss;

[0042] The input parameters and output parameters together constitute a bearing structure-operating condition-performance mapping database, which is used to train an efficient and accurate artificial intelligence model, replacing traditional numerical calculation methods such as empirical formulas or finite element / finite difference methods to predict the performance parameters of tilting pad dynamic oil film bearings;

[0043] Furthermore, the artificial intelligence model includes, but is not limited to: machine learning regression algorithms such as linear regression, ridge regression, support vector machine, Gaussian process regression, random forest, KNN, and neural network regression algorithms such as artificial neural network and deep learning;

[0044] The artificial intelligence reduced-order model can be integrated as a submodule into rotor dynamics analysis, bearing structure parameter optimization or other multi-physics field coupling models.

[0045] Furthermore, the constructed data set is divided into 80% training set and 20% test set, and a deep learning neural network reduced-order model that can replace the thermal elastohydrodynamic bearing model is trained;

[0046] Furthermore, the deep learning neural network reduced-order model is applied to the rotor dynamics analysis and bearing optimization in industrial compressors. According to actual needs, within the recommended bearing structure parameters and operating condition parameter range, the bearing structures with different unknown performances are traversed to quickly and accurately obtain the bearing stiffness, damping and bearing temperature.

[0047] Furthermore, the critical speed and separation margin are calculated based on the obtained stiffness and damping, and the optimal bearing structure that meets the API protocol is quickly screened out.

[0048] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0049] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing based on artificial intelligence, characterized in that: The following steps are involved: Acquire structural parameter data and operating condition parameter data of a target hydrodynamic oil film bearing; The structural parameter data and operating condition parameter data of the target hydrodynamic oil film bearing are input into the trained artificial intelligence model to obtain the performance parameters of the target hydrodynamic oil film bearing.

2. The artificial intelligence-based method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing according to claim 1 is characterized in that: The structural parameters of the target hydrodynamic oil film bearing include: bearing wattage, clearance ratio, aspect ratio, preload and fulcrum type, etc.; The operating parameters of the target dynamic pressure shaft seat include: rotation speed, load, lubrication method, load type, lubricating oil type, oil supply temperature and oil supply pressure.

3. The artificial intelligence-based method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing according to claim 1 is characterized in that: The performance parameters of the target hydrodynamic oil film bearing include mechanical properties and thermodynamic properties; The thermodynamic characteristics include: maximum temperature, average temperature and power loss of the bearing; The mechanical properties include: bearing stiffness, damping and minimum bearing oil film thickness.

4. The artificial intelligence-based method for predicting thermal-mechanical performance of a hydrodynamic oil film bearing according to claim 1, characterized in that: The artificial intelligence model includes a deep learning neural network reduced-order model.

5. The artificial intelligence-based method for predicting the thermal-mechanical performance of a hydrodynamic oil film bearing according to claim 1, characterized in that: The artificial intelligence model is trained through the following steps: Acquire a multi-source data set, wherein the multi-source data set includes hydrodynamic oil film bearing experimental test data, finite element analysis software simulation data, and bearing performance expert knowledge base data; Acquire structural parameter data and operating parameter data of the hydrodynamic oil film bearing from the multi-source data set, and use them as data sets for training the deep learning neural network reduced-order model; The data set is divided into a training set and a test set in a ratio of 4:1, and the training set is input into the deep learning neural network reduced-order model for training to obtain a trained deep learning neural network reduced-order model, and the deep learning neural network reduced-order model outputs the performance parameters of the hydrodynamic oil film bearing.

6. An artificial intelligence-based prediction system for the thermal-mechanical performance of a hydrodynamic oil film bearing, characterized in that: include: A data input module, used for inputting structural parameter data and operating condition parameter data of a target dynamic pressure oil film bearing; The performance prediction output module uses the trained artificial intelligence model and outputs the performance parameters of the target hydrodynamic oil film bearing based on the input structural parameter data and operating condition parameter data.

7. The artificial intelligence-based dynamic oil film bearing thermal-mechanical performance prediction system according to claim 6 is characterized in that: In the data input module, the input structural parameters of the target hydrodynamic oil film bearing include: bearing wattage, clearance ratio, aspect ratio, preload and support type, etc.; The operating parameters of the target dynamic pressure shaft seat include: rotational speed, load, lubrication method, load type, lubricating oil type, oil supply temperature and oil supply pressure.

8. The artificial intelligence-based dynamic oil film bearing thermal-mechanical performance prediction system according to claim 6 is characterized in that: In the performance prediction output module, the output performance parameters of the target hydrodynamic oil film bearing include mechanical characteristics and thermodynamic characteristics; The thermodynamic properties include: the maximum temperature, average temperature and power loss of the bearing; The mechanical properties include: bearing stiffness, damping and minimum bearing oil film thickness.

9. The artificial intelligence-based dynamic oil film bearing thermal-mechanical performance prediction system according to claim 6 is characterized in that: In the performance prediction output module, the artificial intelligence model includes a deep learning neural network reduction model.

10. The artificial intelligence-based dynamic oil film bearing thermal-mechanical performance prediction system according to claim 9, characterized in that: In the performance prediction output module, the artificial intelligence model is trained through the following steps: Acquire a multi-source data set, wherein the multi-source data set includes hydrodynamic oil film bearing experimental test data, finite element analysis software simulation data, and bearing performance expert knowledge base data; Acquire structural parameter data and operating parameter data of the hydrodynamic oil film bearing from the multi-source data set, and use them as data sets for training the deep learning neural network reduced-order model; The data set is divided into a training set and a test set in a ratio of 4:1, and the training set is input into the deep learning neural network reduced-order model for training to obtain a trained deep learning neural network reduced-order model, and the deep learning neural network reduced-order model outputs the performance parameters of the hydrodynamic oil film bearing.

Images