CN-122016316-A - Bearing state monitoring sampling optimization method and system based on reinforcement learning

CN122016316ACN 122016316 ACN122016316 ACN 122016316ACN-122016316-A

Abstract

The invention relates to the technical field of industrial equipment monitoring, in particular to a bearing state monitoring, sampling and optimizing method and a system based on reinforcement learning, which optimize by integrating multidimensional parameters such as equipment operation working condition, real-time vibration signals, bearing key degree and the like, and adopting dynamic baseline prediction, health index correction and sampling strategy, finally, a self-adaptive monitoring sampling strategy which is highly matched with the real-time health state of the bearing is generated, the transition from uniform monitoring to on-demand monitoring is realized, the overall efficiency and the intelligent level of a monitoring system are obviously improved on the premise of ensuring the safety of equipment, and the scientific analysis and the accurate optimization of the bearing state monitoring sampling strategy are realized.

Inventors

LI BAOWEI
LIU SHANGKUN
LIANG FAN
LI HAOYANG

Assignees

河北农业大学

Dates

Publication Date: 20260512
Application Date: 20260129

Claims (10)

1. The bearing state monitoring, sampling and optimizing method is characterized by comprising the following steps: Acquiring equipment operation condition parameters, real-time vibration signals and bearing key degree parameters of a target bearing; generating a health prediction baseline representing the health state of the target bearing according to the equipment operation condition parameters and a preset baseline model, and calculating an initial health index of the target bearing by combining a real-time vibration signal; generating an operation mode feature vector representing the dynamic characteristic of the bearing operation state according to the equipment operation condition parameters, and correcting and compensating the initial health index by combining the bearing key degree parameters to obtain a weighted health index; Judging whether the weighted health index is smaller than a preset health threshold value, if yes, taking a preset safety sampling strategy as a bearing monitoring sampling strategy, and ending the operation; If not, carrying out data combination on the equipment operation condition parameters, the operation mode feature vectors and the weighted health indexes to obtain composite state characterization data; Performing similarity matching on the composite state characterization data and a historical sampling strategy data set, screening historical sampling strategy data with similarity larger than a preset similarity threshold value, and summarizing to obtain an initial sampling strategy set; and inputting the composite state characterization data and the initial sampling strategy set into a pre-constructed bearing sampling strategy optimization model for analysis to obtain a bearing monitoring sampling strategy.
2. The method for monitoring, sampling and optimizing the bearing state according to claim 1, wherein the step of obtaining the equipment operation condition parameter, the real-time vibration signal and the bearing criticality parameter of the target bearing comprises the following steps: setting an initial sampling strategy, wherein the initial sampling strategy at least comprises a sampling time interval and a sampling duration; Acquiring the operation condition parameters and the vibration signals of the target bearing by the multisource sensor according to the initial sampling strategy to obtain the equipment operation condition parameters and the real-time vibration signals of the target bearing; Acquiring a device number corresponding to a target bearing, and matching bearing key degree parameters corresponding to the target bearing in a bearing key degree parameter database according to the device number; The bearing key degree parameter database comprises bearing key degree parameters corresponding to different equipment numbers.
3. The method of claim 1, wherein the calculating the initial health indicator for the target bearing comprises the steps of: inputting the equipment operation condition parameters into a preset final baseline model to perform health baseline prediction, and generating a health prediction baseline representing the health state of the target bearing; Calculating an initial health index of the target bearing according to the health prediction baseline and the real-time vibration signal, wherein the calculation formula is as follows: , Wherein, the Indicating an initial health index of the subject, And The initial sampling end time point and the start time point are respectively represented, And Respectively represent the first and second vibration signals The weights and adjustment parameters corresponding to the individual features, 、 And Respectively represent time of day First, the Measured values, baseline predicted values and preset standard deviations corresponding to the features, Representing the total number of vibration signal features.
4. The method of claim 1, wherein the obtaining composite state characterization data comprises the steps of: The method comprises the steps of constructing a working condition-operation mode mapping relation, wherein the working condition-operation mode mapping relation is used for mapping continuous equipment operation working condition parameters into an operation mode feature vector representing dynamic characteristics of the bearing operation state; generating an operation mode feature vector representing the dynamic characteristic of the bearing operation state according to the equipment operation condition parameter and the condition-operation mode mapping relation; the operation mode feature vector comprises the intensity that the bearing operation state belongs to various preset operation modes; Calculating a real-time risk coefficient of the target bearing according to the operation mode feature vector; Correcting and compensating the initial health index according to the real-time risk coefficient and the bearing key degree parameter to obtain a weighted health index; judging whether the weighted health index is smaller than a preset health threshold value or not; If yes, taking a preset safety sampling strategy as a bearing monitoring sampling strategy, and ending the operation; if not, the equipment operation condition parameters, the operation mode feature vectors and the weighted health indexes are subjected to data combination to obtain the composite state representation data.
5. The method of claim 1, wherein the obtaining an initial sampling strategy set comprises the steps of: Acquiring historical sampling strategy data of a plurality of bearings in different running states through the Internet of things to obtain a historical sampling strategy data set, wherein the historical sampling strategy data comprises historical composite state representation data of the different bearings in the different running states and a adopted historical bearing monitoring sampling strategy; Calculating the similarity of the composite state characterization data and each historical sampling strategy data in the historical sampling strategy data set; And if the similarity between the composite state characterization data and each historical sampling strategy data in the historical sampling strategy data set is larger than a preset similarity threshold, the historical sampling strategy data is reserved, and all the historical bearing monitoring sampling strategies contained in the reserved historical sampling strategy data are summarized to obtain an initial sampling strategy set.
6. The method of claim 1, wherein the obtaining a bearing condition monitoring sampling strategy comprises the steps of: Sequentially calculating the similarity of each historical sampling strategy data and other historical sampling strategy data in the historical sampling strategy data set, and distributing a historical bearing monitoring sampling strategy contained in the historical sampling strategy data with the similarity larger than the preset similarity threshold value for each historical sampling strategy data as a historical initial sampling strategy set corresponding to the historical sampling strategy data; constructing a final bearing sampling strategy optimization model according to the historical initial sampling strategy set corresponding to each historical sampling strategy data in the historical sampling strategy data set; And inputting the composite state characterization data and the initial sampling strategy set into a pre-constructed final bearing sampling strategy optimization model for analysis to obtain a bearing monitoring sampling strategy.
7. The method of claim 6, wherein the final bearing sampling strategy optimization model employs a dual depth Q network model.
8. A system for implementing the bearing condition monitoring sampling optimization method of any one of claims 1-7, comprising: the data acquisition module is used for acquiring equipment operation condition parameters, real-time vibration signals and bearing key degree parameters of the target bearing; the initial health index metering module is used for generating a health prediction baseline representing the health state of the target bearing according to the equipment operation condition parameters and a preset baseline model, and calculating an initial health index of the target bearing by combining a real-time vibration signal; The weighted health index analysis module is used for generating an operation mode feature vector representing the dynamic characteristic of the operation state of the bearing according to the operation condition parameters of the equipment, correcting and compensating the initial health index by combining the bearing key degree parameters to obtain a weighted health index, judging whether the weighted health index is smaller than a preset health threshold value, if so, taking a preset safety sampling strategy as a bearing monitoring sampling strategy, and ending the operation; The initial sampling strategy matching module is used for carrying out similarity matching on the composite state characterization data and the historical sampling strategy data set, screening historical sampling strategy data with similarity larger than a preset similarity threshold value, and summarizing to obtain an initial sampling strategy set; And the bearing monitoring sampling strategy optimization module is used for inputting the composite state characterization data and the initial sampling strategy set into a pre-constructed bearing sampling strategy optimization model for analysis to obtain a bearing monitoring sampling strategy.
9. An electronic device comprising a memory and a processor, wherein the memory is configured to store computer-executable instructions and the processor is configured to execute the computer-executable instructions, which when executed by the processor, implement the steps of the method according to any one of claims 1-7.
10. A computer storage medium having stored thereon computer executable instructions which when executed by a processor perform the steps of the method according to any of claims 1-7.

Description

Bearing state monitoring sampling optimization method and system based on reinforcement learning Technical Field The invention relates to the technical field of industrial equipment monitoring, in particular to a bearing state monitoring sampling optimization method and system based on reinforcement learning. Background The bearing is used as a core component of the rotary machine, and the running state of the bearing is directly related to the safety, stability and efficiency of the whole equipment and even the production line. Therefore, the method for monitoring the state and managing the health of the bearing and implementing predictive maintenance is a key technology for guaranteeing the reliability of equipment and reducing the operation and maintenance cost in the industrial field. In the prior art, the dynamic influence of the change of the operating condition on the vibration baseline is not considered when the health state of the bearing is evaluated, so that misjudgment of the health state of the same bearing is easy to occur under different working conditions, and meanwhile, the key degree difference of the bearing in the equipment is not included in the evaluation process, so that the evaluation result cannot reflect the real maintenance priority. The isolated and static evaluation mode leads to the failure of the monitoring strategy to carry out self-adaptive adjustment according to the accurate and quantized result of the health state, thereby causing resource waste and insufficient risk management and control. Disclosure of Invention Aiming at the problems in the related art, the invention provides a bearing state monitoring sampling optimization method and a system based on reinforcement learning, so as to overcome the technical problems in the prior art. In order to solve the technical problems, the invention provides a bearing state monitoring sampling optimization method based on reinforcement learning, which comprises the following steps: Acquiring equipment operation condition parameters, real-time vibration signals and bearing key degree parameters of a target bearing; generating a health prediction baseline representing the health state of the target bearing according to the equipment operation condition parameters and a preset baseline model, and calculating an initial health index of the target bearing by combining a real-time vibration signal; generating an operation mode feature vector representing the dynamic characteristic of the bearing operation state according to the equipment operation condition parameters, and correcting and compensating the initial health index by combining the bearing key degree parameters to obtain a weighted health index; Judging whether the weighted health index is smaller than a preset health threshold value, if yes, taking a preset safety sampling strategy as a bearing monitoring sampling strategy, and ending the operation; If not, carrying out data combination on the equipment operation condition parameters, the operation mode feature vectors and the weighted health indexes to obtain composite state characterization data; Performing similarity matching on the composite state characterization data and a historical sampling strategy data set, screening historical sampling strategy data with similarity larger than a preset similarity threshold value, and summarizing to obtain an initial sampling strategy set; and inputting the composite state characterization data and the initial sampling strategy set into a pre-constructed bearing sampling strategy optimization model for analysis to obtain a bearing monitoring sampling strategy. Preferably, the specific steps for acquiring the equipment operation condition parameters, the real-time vibration signals and the bearing key degree parameters of the target bearing are as follows: setting an initial sampling strategy, wherein the initial sampling strategy at least comprises a sampling time interval and a sampling duration; Acquiring the operation condition parameters and the vibration signals of the target bearing by the multisource sensor according to the initial sampling strategy to obtain the equipment operation condition parameters and the real-time vibration signals of the target bearing; Acquiring a device number corresponding to a target bearing, and matching bearing key degree parameters corresponding to the target bearing in a bearing key degree parameter database according to the device number; The bearing key degree parameter database comprises bearing key degree parameters corresponding to different equipment numbers. Preferably, the specific steps of generating a health prediction baseline representing the health state of the target bearing according to the equipment operation condition parameters and a preset baseline model, and calculating an initial health index of the target bearing by combining a real-time vibration signal are as follows: S21, inputting the equipment operation condition pa