Lubricating oil aging prediction method based on environmental differences

Abstract

This invention discloses a method for predicting lubricating oil aging based on environmental differential aging, relating to the field of industrial maintenance. The method includes: collecting initial performance parameters of lubricating oil under initial environmental conditions to construct an initial lubricating oil dataset; collecting usage environmental conditions and performance parameters at preset sliding time window nodes to construct a lubricating oil usage dataset; using the initial lubricating oil dataset as static input and the usage dataset as dynamic input, constructing a lubricating oil environmental differential aging prediction model based on a long short-term memory network; inputting the desired usage environmental conditions into the lubricating oil environmental differential aging prediction model to predict lubricating oil aging, generating and outputting the lubricating oil aging prediction result. This method solves the technical problem of insufficient accuracy in existing lubricating oil aging prediction methods, achieving the technical effect of improving the accuracy of lubricating oil aging prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial maintenance, and particularly relates to a lubricating oil aging prediction method based on environment difference. BACKGROUND

[0002] Lubricating oil plays a role of lubrication, cooling, rust prevention and cleaning in mechanical equipment. However, with the increase of use time, the lubricating oil will gradually age, and after the lubricating oil ages, its performance parameters such as viscosity, flash point, kinematic viscosity and base value will change, resulting in reduced lubrication effect and increased energy consumption. By predicting the aging of the lubricating oil, the performance problems of the lubricating oil can be found and handled in time, and the performance of the lubricating oil can be maintained, thereby reducing the wear of the equipment. In the prior art, the aging prediction of the lubricating oil is usually based on simple physical and chemical tests or statistical models based on historical data, and these methods do not fully consider the changes of the performance of the lubricating oil under different environmental conditions. In actual use, the performance of the lubricating oil is affected by various environmental factors such as temperature, humidity and pressure. Since the prior art lacks accurate capture and processing of changes in environmental factors, the prediction results are often not accurate enough.

[0003] In the related art at present, the lubricating oil aging prediction method has the technical problem of inaccurate prediction results. SUMMARY

[0004] The present application provides a lubricating oil aging prediction method based on environment difference, which adopts the technical means of continuously collecting the use environment conditions and corresponding use performance parameters of the lubricating oil on a preset sliding time window node, constructing a lubricating oil environment difference aging prediction model, and predicting the aging of the lubricating oil, so that the lubricating oil environment difference aging prediction model learns the change rule of the performance of the lubricating oil under different environmental conditions, thereby accurately predicting the performance change of the lubricating oil under different environmental conditions, and achieving the technical effect of improving the accuracy of the lubricating oil aging prediction.

[0005] The present application provides a lubricating oil aging prediction method based on environment difference, which includes:

[0006] interactively collecting initial performance parameters of the lubricating oil under initial environment conditions, and constructing an initial data set of the lubricating oil;

[0007] based on the initial environment conditions, collecting use environment conditions and corresponding use performance parameters of the lubricating oil on a preset sliding time window node, and constructing a use data set of the lubricating oil, wherein the use environment conditions and the initial environment conditions contain the same types of environmental variables, and the numerical values of the environmental variables change with time, and the use performance parameters and the initial performance parameters contain the same types of parameters;

[0008] The initial data set of the lubricating oil is taken as static input, the lubricating oil use data set is taken as dynamic input, and a lubricating oil environment differential aging prediction model is constructed based on a long short-term memory network;

[0009] The lubricating oil aging prediction result is outputted.

[0010] In a possible implementation, the lubricating oil initial data set is taken as static input, the lubricating oil use data set is taken as dynamic input, and a lubricating oil environment differential aging prediction model is constructed based on a long short-term memory network, and the following processing is performed:

[0011] The initial data set of the lubricating oil is taken as an initial embedding vector for static input, the use environment condition in the lubricating oil use data set is taken as dynamic input, and the corresponding use performance parameter is taken as output information, and a lubricating oil use performance parameter predictor is constructed based on long short-term memory network training;

[0012] The use environment condition and the use performance parameter in the lubricating oil use data set are taken as input information, and the remaining use life of the lubricating oil is taken as output information, and a lubricating oil remaining use life predictor is constructed based on long short-term memory network training;

[0013] The lubricating oil use performance parameter predictor and the lubricating oil remaining use life predictor are sequentially connected to construct a lubricating oil environment differential aging prediction model.

[0014] In a possible implementation, the use environment condition in the lubricating oil use data set is taken as dynamic input, and the following processing is performed:

[0015] The use environment condition includes at least three environment variables, any two environment variables in the use environment condition are taken as target variables, and the environment variables excluding the target variables in the use environment condition and the use performance parameter are taken as control variables, a partial correlation analysis is performed, and a partial correlation coefficient between the two target variables is obtained;

[0016] If the partial correlation coefficient is greater than a preset partial correlation threshold, the two target variables are taken as new dynamic input after being interacted.

[0017] In a possible implementation, the lubricating oil aging prediction result is outputted by inputting the use environment condition into the lubricating oil environment differential aging prediction model, and the following processing is performed:

[0018] The use environment condition is inputted into the lubricating oil use performance parameter predictor, and a predicted use performance parameter is outputted.

[0019] traverse the predicted use performance parameters, extract a first predicted use performance parameter;

[0020] determine whether the first predicted use performance parameter exceeds a first predicted use performance parameter threshold, if exceeding, activate a lubricating oil aging warning signal, generate a lubricating oil aging prediction result based on the lubricating oil aging warning signal and output.

[0021] In a possible implementation, the traverse the predicted use performance parameters, the following processing is performed:

[0022] If none of the predicted use performance parameters exceeds the corresponding predicted use performance parameter threshold, input the to-be-used environmental conditions and the predicted use performance parameters into the lubricating oil remaining use life predictor to perform lubricating oil remaining use life prediction, generate a lubricating oil aging prediction result based on the lubricating oil remaining use life prediction result and output.

[0023] In a possible implementation, the input the to-be-used environmental conditions and the predicted use performance parameters into the lubricating oil remaining use life predictor to perform lubricating oil remaining use life prediction, the following processing is performed:

[0024] perform time series on the to-be-used environmental conditions and the predicted use performance parameters, to obtain to-be-used environmental condition time series data and predicted use performance parameter time series data;

[0025] based on the to-be-used environmental condition time series data, randomly extract first to-be-used environmental variable time series data;

[0026] based on the predicted use performance parameter time series data, randomly extract first predicted use performance parameter time series data;

[0027] interact the first to-be-used environmental variable time series data and the first predicted use performance parameter time series data, to obtain random interaction time series data;

[0028] perform time trend grabbing on the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data, and perform lubricating oil remaining use life prediction based on the trend direction.

[0029] In a possible implementation, the perform time trend grabbing on the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data, and perform lubricating oil remaining use life prediction based on the trend direction, the following processing is performed:

[0030] extracting trend components, seasonal components and random components in the to-be-used environment condition time series data, the predicted use performance parameter time series data and the random interaction time series data;

[0031] characteristic capturing is respectively performed on the trend components, the seasonal components and the random components to obtain trend component characteristics, seasonal component characteristics and random component characteristics;

[0032] According to the trend component characteristics, seasonal component characteristics and random component characteristics, combined with weight distribution, the residual service life of the lubricating oil is predicted.

[0033] In a possible implementation, the lubricating oil aging prediction by inputting the to-be-used environment condition into the lubricating oil environment differential aging prediction model, generating and outputting a lubricating oil aging prediction result, performs the following processing:

[0034] data augmentation is performed on the lubricating oil use data set to obtain an augmented lubricating oil use data set;

[0035] The augmented lubricating oil use data set is used to test the lubricating oil environment differential aging prediction model;

[0036] According to the test result, the lubricating oil environment differential aging prediction model is corrected.

[0037] The method for predicting the aging of lubricating oil based on environmental differences provided in the present application is as follows: first, the initial performance parameters of the lubricating oil under the initial environmental conditions are collected and interacted to construct an initial data set of the lubricating oil; then, based on the initial environmental conditions, the use environment conditions and the corresponding use performance parameters of the lubricating oil are collected and obtained at the preset sliding time window nodes to construct a use data set of the lubricating oil, wherein the use environment conditions and the initial environment conditions contain the same types of environmental variables, and the numerical values of the environmental variables change with time; the use performance parameters and the initial performance parameters contain the same types of parameters; then, the initial data set of the lubricating oil is taken as a static input, the use data set of the lubricating oil is taken as a dynamic input, a long short-term memory network is used to construct a lubricating oil environment differential aging prediction model, and finally, the to-be-used environment condition is input into the lubricating oil environment differential aging prediction model to predict the aging of the lubricating oil, generate and output a lubricating oil aging prediction result, thereby achieving the technical effect of improving the accuracy of the lubricating oil aging prediction. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application will be briefly introduced below. In the present application, flowcharts are used to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously according to needs. Meanwhile, other operations can be added to these processes, or one or more steps of operations can be removed from these processes.

[0039] Figure 1 A flowchart of the lubricating oil aging prediction method based on environmental difference provided for the embodiments of the present application is shown in the figure.

[0040] Figure 2 A flowchart of the lubricating oil remaining service life prediction of the lubricating oil aging prediction method based on environmental difference provided for the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0041] The foregoing description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the embodiments of the present application can be implemented according to the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described as follows.

[0042] In order to make the purposes, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings, and the described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of protection of the present application.

[0043] In the following description, "some embodiments" are described, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict, and the term "first\second" referred to only distinguishes similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art in the technical field of the present application. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0044] The embodiments of the present application provide a lubricating oil aging prediction method based on environmental difference, as shown in the figure. Figure 1As shown, the method comprises:

[0045] In step S100, initial performance parameters of the lubricating oil under initial environmental conditions are collected, and an initial data set of the lubricating oil is constructed. Specifically, the initial environmental conditions are the environmental conditions of the lubricating oil when it is not used or just put into use, and the initial performance parameters are the performance parameters exhibited by the lubricating oil under the initial environmental conditions, which are used to describe the initial state of the lubricating oil. The environmental parameters of the lubricating oil in the initial state, such as temperature, humidity, and pressure, are determined, and the initial performance parameters of the lubricating oil under the initial environmental conditions, such as viscosity, acid value, and flash point, are collected using professional testing equipment or methods. The collected initial environmental conditions and corresponding initial performance parameters are arranged in the form of a data set to form the initial data set of the lubricating oil.

[0046] In step S200, based on the initial environmental conditions, the use environmental conditions and corresponding use performance parameters of the lubricating oil are collected at preset sliding time window nodes to construct a use data set of the lubricating oil, wherein the use environmental conditions and the initial environmental conditions contain the same types of environmental variables, and the numerical values of the environmental variables change over time, and the use performance parameters and the initial performance parameters contain the same types of parameters. Specifically, the preset sliding time window is a preset time period for regularly collecting the use data of the lubricating oil, and the preset sliding time window is set according to actual needs. At each preset sliding time window node, the use environmental conditions of the lubricating oil (the environmental conditions of the lubricating oil in the actual use process) are collected, and the types of environmental variables of the collected use environmental conditions are the same as those of the initial environmental conditions, i.e., if the initial environmental conditions include temperature, humidity, and pressure, the collected use environmental conditions also include the numerical values of temperature, humidity, and pressure. At each time window node, the use performance parameters of the lubricating oil under the corresponding use environmental conditions are collected, the use performance parameters are the performance parameters exhibited by the lubricating oil in the actual use process, which are used to describe the state change of the lubricating oil, and the types of use performance parameters are also the same as those of the initial performance parameters. The collected use environmental conditions and corresponding use performance parameters are arranged in the form of a data set to form the use data set of the lubricating oil.

[0047] In step S300, the initial data set of the lubricating oil is taken as static input, the lubricating oil use data set is taken as dynamic input, and a lubricating oil environment differential aging prediction model is constructed based on a long short-term memory network. Specifically, the initial data set of the lubricating oil and the use data set of the lubricating oil are subjected to necessary preprocessing, such as data cleaning, normalization and the like, to meet the requirements of model training. By utilizing the ability of the long short-term memory network to process time series data, a lubricating oil environment differential aging prediction model is constructed, the initial data set of the lubricating oil is taken as static input, the use data set of the lubricating oil is taken as dynamic input, and the long short-term memory network is trained so that it can learn the aging law of the lubricating oil under different environmental conditions.

[0048] In a possible implementation, step S300 further includes step S310, inputting the initial data set of the lubricating oil as an initial embedding vector, inputting the use environment condition in the use data set of the lubricating oil as a dynamic input, inputting the corresponding use performance parameter as output information, and constructing a lubricating oil use performance parameter predictor based on long short-term memory network training. Specifically, the initial data set of the lubricating oil is converted into an initial embedding vector by a word embedding technology or the like, a long short-term memory network structure is designed, the initial embedding vector is input as a static input, the use environment condition is input as a dynamic input (changes over time), and the use performance parameter corresponding to the use environment condition is input as output information. The lubricating oil use performance parameter predictor is trained using the use data set of the lubricating oil, so that the long short-term memory network can learn the correlation between the environment condition and the lubricating oil performance parameter. In step S320, the use environment condition and the use performance parameter in the use data set of the lubricating oil are input as input information, the remaining use life of the lubricating oil is input as output information, and a lubricating oil remaining use life predictor is constructed based on long short-term memory network training. Specifically, the use environment condition and the use performance parameter in the use data set of the lubricating oil are used as input information, and a long short-term memory network structure is designed again, so that the use environment condition and the use performance parameter are input, and the remaining use life of the lubricating oil (the time or mileage that the lubricating oil can continue to use under a specific condition) is output. The lubricating oil remaining use life predictor is trained using the use data set of the lubricating oil, so that the long short-term memory network can predict the remaining use life of the lubricating oil. In step S330, the lubricating oil use performance parameter predictor and the lubricating oil remaining use life predictor are sequentially connected to construct a lubricating oil environment differential aging prediction model. Specifically, the lubricating oil use performance parameter predictor trained in step S310 and the lubricating oil remaining use life predictor trained in step S320 are sequentially connected to form a complete lubricating oil environment differential aging prediction model. This implementation decomposes the aging prediction of the lubricating oil into two subtasks, first predicts the use performance parameter of the lubricating oil under different use environment conditions, and then predicts the remaining use life of the lubricating oil based on the use performance parameter. This divide-and-conquer strategy helps the lubricating oil environment differential aging prediction model to more accurately learn the complex process of lubricating oil aging, and achieves the technical effect of more accurately predicting the aging of the lubricating oil.

[0049] In a possible implementation, the step S310 of taking the use environment conditions in the lubricating oil use dataset as dynamic inputs further includes a step S311. The use environment conditions include at least three environment variables, any two of the use environment conditions are taken as target variables, and the environment variables excluding the target variables in the use environment conditions and the use performance parameter are taken as control variables to perform a partial correlation analysis to obtain a partial correlation coefficient between the two target variables. Specifically, the use environment condition data is extracted from the lubricating oil use dataset. The use environment conditions include at least three environment variables such as temperature, humidity, pressure, etc. Any two of all the environment variables are selected as target variables, for example, temperature and humidity are selected as a group of target variables. The environment variables other than the selected target variables and the use performance parameter are taken as control variables, which are used to eliminate the potential influence of the control variables on the relationship between the target variables in the partial correlation analysis. A statistical method such as a partial correlation coefficient formula or statistical software is used to calculate the partial correlation coefficient between the two target variables. The partial correlation coefficient measures the correlation between the two target variables while controlling other variables. In step S312, if the partial correlation coefficient is greater than a preset partial correlation threshold, the two target variables are interacted to generate a new dynamic input. Specifically, the calculated partial correlation coefficient is compared with the preset partial correlation threshold, which is determined according to the actual application background and statistical theory and is used to determine whether there is a significant correlation between the two target variables. If the partial correlation coefficient is greater than the preset partial correlation threshold, it indicates that there is a significant correlation between the two target variables after controlling other variables. In this case, the two target variables are interacted (such as multiplication, division, polynomial, etc.) to generate a new variable as a new dynamic input. The interaction processing is used to further capture the nonlinear relationship between the two target variables. The generated new dynamic input is added to the original dynamic input set for training of the long short-term memory network. This implementation identifies, through the partial correlation analysis, which environment variables have a significant correlation under the use environment conditions. These correlations may have an important influence on the use performance parameter of the lubricating oil. The environment variables with significant correlations are interacted to generate a new dynamic input, which helps the long short-term memory network model to more accurately capture the interaction between the environment variables, thereby achieving the technical effect of improving the accuracy of the lubricating oil use performance parameter prediction.

[0050] In step S400, the environment conditions to be used are input into the lubricating oil environment differential aging prediction model for lubricating oil aging prediction, and a lubricating oil aging prediction result is generated and output. Specifically, the environment conditions in which the lubricating oil to be subjected to lubricating oil aging prediction is located are input into the trained lubricating oil environment differential aging prediction model. The lubricating oil environment differential aging prediction model performs lubricating oil aging prediction according to the input environment conditions to be used, in combination with the learned lubricating oil aging law, and outputs the prediction result (predicted lubricating oil aging condition or performance index) in the form of a numerical value, a chart, or the like for reference and decision-making by a user. The embodiments of the present application use technical means such as continuously collecting the use environment conditions and corresponding use performance parameters of the lubricating oil on the preset sliding time window node, constructing a lubricating oil environment differential aging prediction model, and performing lubricating oil aging prediction, so that the lubricating oil environment differential aging prediction model learns the change law of the lubricating oil performance under different environment conditions, thereby accurately predicting the performance change of the lubricating oil under different environment conditions, and achieving the technical effect of improving the accuracy of lubricating oil aging prediction.

[0051] In a possible implementation, step S400 further includes step S410: inputting the to-be-used environmental condition into the lubricating oil use performance parameter predictor, and outputting the predicted use performance parameter. Specifically, the actual environmental condition in which the lubricating oil to be predicted is located is input into the trained lubricating oil use performance parameter predictor, the lubricating oil use performance parameter predictor predicts the use performance parameter of the lubricating oil under the environmental condition based on the input to-be-used environmental condition, and outputs the predicted use performance parameter as a prediction result. Step S420: traversing the predicted use performance parameter, and extracting a first predicted use performance parameter. Specifically, the predicted use performance parameter obtained in step S410 is traversed, and the first (or specified order) predicted use performance parameter is selected as the first predicted use performance parameter. Step S430: determining whether the first predicted use performance parameter exceeds a first predicted use performance parameter threshold, and if so, activating a lubricating oil aging warning signal, generating a lubricating oil aging prediction result based on the lubricating oil aging warning signal, and outputting the lubricating oil aging prediction result. Specifically, the first predicted use performance parameter is compared with a preset first predicted use performance parameter threshold, the first predicted use performance parameter threshold is a preset numerical value, and is used to determine whether the first predicted use performance parameter exceeds a normal range. If the first predicted use performance parameter exceeds the first predicted use performance parameter threshold, it indicates that the performance of the lubricating oil has exceeded the safe use range, and the lubricating oil aging warning signal is activated at this time. Based on the activated lubricating oil aging warning signal, the lubricating oil aging prediction result is generated, and the result is output, the lubricating oil aging prediction result includes the performance state (aging warning) of the lubricating oil and related performance index data. This implementation predicts the use performance parameter of the lubricating oil under different use environmental conditions, monitors the working state of the lubricating oil in real time, and timely issues a warning when the performance problem is predicted, so that the technical effect of avoiding equipment failure or damage caused by lubricating oil aging is achieved.

[0052] In a possible implementation, the step of traversing the predicted use performance parameters, S400, further comprises a step S440 of inputting the to-be-used environmental conditions and the predicted use performance parameters into the lubricating oil residual use life predictor to predict the lubricating oil residual use life, generating and outputting the lubricating oil aging prediction result based on the lubricating oil residual use life prediction result, if none of the predicted use performance parameters exceeds the corresponding predicted use performance parameter threshold. Specifically, after traversing the predicted use performance parameters, it is checked whether any of the predicted use performance parameters exceeds the corresponding predicted use performance parameter threshold (a numerical value set for each predicted use performance parameter to determine whether the predicted use performance parameter exceeds the normal range). If none of the predicted use performance parameters exceeds the corresponding predicted use performance parameter threshold, it indicates that the performance of the current lubricating oil is within the normal range. At this time, the to-be-used environmental conditions and the predicted use performance parameters are input into the lubricating oil residual use life predictor as input, and the lubricating oil residual use life predictor predicts the residual use life of the lubricating oil under the environmental conditions based on the input to-be-used environmental conditions and the predicted use performance parameters. Based on the lubricating oil residual use life prediction result output by the lubricating oil residual use life predictor, the lubricating oil aging prediction result is generated, which includes the residual use life information and the possible aging trend or state of the lubricating oil, and finally the generated lubricating oil aging prediction result is output to the user or the system. This implementation further predicts the lubricating oil residual use life when all the predicted use performance parameters are within the normal range, and provides more comprehensive and accurate lubricating oil aging prediction information for the user or the system, thereby achieving the technical effect of helping the user or the system to understand the residual use life of the lubricating oil in advance, so as to formulate a suitable maintenance plan or replacement strategy.

[0053] As Figure 2In a possible implementation, as shown, the step S440 of inputting the to-be-used environmental condition and the predicted use performance parameter into the lubricating oil residual use life predictor to predict the lubricating oil residual use life further includes steps S441 to S445. Specifically, the step S441 is to perform time series on the to-be-used environmental condition and the predicted use performance parameter to obtain to-be-used environmental condition time series data and predicted use performance parameter time series data. Specifically, time series is performed on the to-be-used environmental condition (such as data of temperature, humidity, pressure, etc. changing over time) to generate to-be-used environmental condition time series data. Similarly, time series is also performed on the predicted use performance parameter (such as data of viscosity, acid value, etc. changing over time) to generate predicted use performance parameter time series data. The step S442 is to randomly extract first to-be-used environmental variable time series data based on the to-be-used environmental condition time series data. Specifically, time series data of an environmental variable (such as temperature or humidity) is randomly selected from the to-be-used environmental condition time series data. The step S443 is to randomly extract first predicted use performance parameter time series data based on the predicted use performance parameter time series data. Specifically, time series data of a performance parameter (such as viscosity or acid value) is randomly selected from the predicted use performance parameter time series data. The step S444 is to interact the first to-be-used environmental variable time series data and the first predicted use performance parameter time series data to obtain random interaction time series data. Specifically, the first to-be-used environmental variable time series data and the first predicted use performance parameter time series data randomly extracted are interacted, and the interaction can be simple multiplication or addition, or more complex mathematical or statistical operation, to explore the potential relationship between the environmental variable and the performance parameter. The step S445 is to perform time trend extraction on the to-be-used environmental condition time series data, the predicted use performance parameter time series data, and the random interaction time series data to predict the lubricating oil residual use life based on the trend. Specifically, time trend analysis is performed on the to-be-used environmental condition time series data, the predicted use performance parameter time series data, and the random interaction time series data, time series analysis techniques (such as moving average, exponential smoothing, etc.) are used to identify the trend of these time series data, and a prediction model is used to predict the lubricating oil residual use life based on the trend. This implementation comprehensively and deeply analyzes the law of change of the lubricating oil performance over time by performing time series, random extraction, interaction, and time trend extraction on the to-be-used environmental condition and the predicted use performance parameter, and the random extraction and interaction analysis increase the diversity and robustness of the lubricating oil residual use life predictor, so that the lubricating oil residual use life predictor can adapt to changes under different environments and conditions, and the technical effects of improving the accuracy and reliability of the lubricating oil residual use life prediction are achieved.

[0054] In a possible implementation, step S445 further includes step S4451 of extracting trend components, seasonal components and random components from the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data. Specifically, the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data are decomposed into three main components, i.e., trend components (long-term trend), seasonal components (periodic change) and random components (random fluctuation or error), by using a time series decomposition technique such as STL decomposition, X-12-ARIMA seasonal adjustment or the like. Among them, the trend component represents the long-term trend of the data changing over time, such as rising, falling or stable; the seasonal component represents the periodic change existing in the data, which is related to seasonal or periodic events; and the random component represents the random fluctuation or error in the data, which is caused by unpredictable factors. Step S4452 includes capturing characteristics of the trend components, the seasonal components and the random components respectively to obtain trend component characteristics, seasonal component characteristics and random component characteristics. Specifically, characteristics of each component (trend component, seasonal component and random component) are captured, and the characteristics of the trend component include a slope, a growth rate or a decline rate and the like; the characteristics of the seasonal component include a seasonal amplitude, a seasonal period and the like; and the characteristics of the random component include volatility, skewness and kurtosis and the like. Step S4453 includes performing lubricating oil residual use life prediction according to the trend component characteristics, the seasonal component characteristics and the random component characteristics in combination with weight distribution. Specifically, the captured trend component characteristics, the seasonal component characteristics and the random component characteristics are combined with preset or learned weights to perform weighted combination, and the result of the weighted combination is used to perform lubricating oil residual use life prediction. This implementation decomposes the original data into trend components, seasonal components and random components by using a time series decomposition technique, and captures characteristics of these components respectively, which more accurately understands and analyzes different change patterns in the data, more accurately captures and simulates the influence of these components on the lubricating oil residual use life, and thus achieves the technical effect of improving the prediction accuracy.

[0055] In a possible implementation, step S400 further includes step S450, data augmentation is performed on the lubricating oil use dataset to obtain an augmented lubricating oil use dataset. Specifically, a part of representative data is selected from the original lubricating oil use dataset, which covers different use environment conditions, use time length, lubricating oil types, etc. Transformations are performed on the selected data, including rotation, translation, scaling, noise addition, label smoothing, etc., to increase the diversity and richness of the dataset, and new lubricating oil use performance parameter data is simulated and generated by using existing physical models or statistical models. The transformed data and the simulated and generated data are fused with the original lubricating oil use dataset to form the augmented lubricating oil use dataset. In step S460, the lubricating oil environment differential aging prediction model is tested using the augmented lubricating oil use dataset. Specifically, the trained lubricating oil environment differential aging prediction model is loaded, the lubricating oil environment differential aging prediction model is tested using the augmented lubricating oil use dataset, the performance of the lubricating oil environment differential aging prediction model on the augmented lubricating oil use dataset is evaluated, and the evaluation indexes can include accuracy, recall rate, F1 score, mean square error, etc. In step S470, the lubricating oil environment differential aging prediction model is corrected according to the test result. Specifically, the performance of the lubricating oil environment differential aging prediction model on the augmented lubricating oil use dataset is analyzed, the deficiencies of the lubricating oil environment differential aging prediction model are found out, the lubricating oil environment differential aging prediction model is adjusted according to the performance analysis result, including changing the model structure, adjusting the model parameters, introducing new features, etc. The model structure and parameters after adjustment are used for retraining on the augmented lubricating oil use dataset, and the above steps are repeated until the performance of the lubricating oil environment differential aging prediction model on the augmented lubricating oil use dataset reaches the expected level. This implementation effectively improves the performance of the lubricating oil environment differential aging prediction model through data augmentation, model testing and correction, etc., and achieves the technical effect that the lubricating oil environment differential aging prediction model can more accurately and reliably predict the remaining service life of the lubricating oil and the aging trend under the environmental conditions.

[0056] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application. In some cases, the actions or steps described in the present application can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A method for predicting the aging of lubricating oil based on environmental differences, characterized by, The method comprises: interactively collecting initial performance parameters of lubricating oil under initial environmental conditions to construct an initial data set of lubricating oil; based on the initial environmental conditions, collecting usage environmental conditions and corresponding usage performance parameters of lubricating oil on a preset sliding time window node to construct a usage data set of lubricating oil, wherein the usage environmental conditions and the initial environmental conditions contain the same types of environmental variables, and the numerical values of the environmental variables change over time, and the usage performance parameters and the initial performance parameters contain the same types of parameters; inputting the initial data set of lubricating oil as static input and the usage data set of lubricating oil as dynamic input, and constructing a lubricating oil environmental differential aging prediction model based on a long short-term memory network; inputting the to-be-used environmental conditions into the lubricating oil environmental differential aging prediction model for lubricating oil aging prediction, generating a lubricating oil aging prediction result, and outputting the lubricating oil aging prediction result; wherein the inputting the initial data set of lubricating oil as static input and the usage data set of lubricating oil as dynamic input, and constructing a lubricating oil environmental differential aging prediction model based on a long short-term memory network, comprises: inputting the initial data set of lubricating oil as an initial embedding vector for static input, inputting the usage environmental conditions in the usage data set of lubricating oil as dynamic input, and inputting the corresponding usage performance parameters as output information, and constructing a lubricating oil usage performance parameter predictor based on a long short-term memory network training; inputting the usage environmental conditions and the usage performance parameters in the usage data set of lubricating oil as input information, and inputting the remaining usage life of lubricating oil as output information, and constructing a lubricating oil remaining usage life predictor based on a long short-term memory network training; sequentially connecting the lubricating oil usage performance parameter predictor and the lubricating oil remaining usage life predictor to construct a lubricating oil environmental differential aging prediction model; wherein the inputting the usage environmental conditions in the usage data set of lubricating oil as dynamic input comprises: the usage environmental conditions include at least three environmental variables, any two environmental variables in the usage environmental conditions are taken as target variables, the environmental variables excluding the target variables in the usage environmental conditions and the usage performance parameters are taken as control variables, a partial correlation analysis is performed to obtain a partial correlation coefficient between the two target variables; if the partial correlation coefficient is greater than a preset partial correlation threshold, the two target variables are interacted to serve as new dynamic input.

2. The environment difference-based lubricating oil aging prediction method according to claim 1, characterized by, The method of inputting the to-be-used environmental conditions into the lubricating oil environmental differential aging prediction model for aging prediction, generating a lubricating oil aging prediction result, and outputting the lubricating oil aging prediction result, comprises: inputting the to-be-used environmental conditions into the lubricating oil usage performance parameter predictor to output a predicted usage performance parameter; traversing the predicted usage performance parameter to extract a first predicted usage performance parameter; judging whether the first predicted usage performance parameter exceeds a first predicted usage performance parameter threshold, if yes, activating a lubricating oil aging warning signal, and generating a lubricating oil aging prediction result based on the lubricating oil aging warning signal and outputting the lubricating oil aging prediction result.

3. The environment difference-based lubricating oil aging prediction method according to claim 2, characterized by, The method of traversing the predicted usage performance parameter further comprises: If none of the predicted use performance parameters exceeds the corresponding predicted use performance parameter threshold, input the to-be-used environmental condition and the predicted use performance parameter into the lubricating oil remaining use life predictor to perform lubricating oil remaining use life prediction, and generate and output a lubricating oil aging prediction result based on a lubricating oil remaining use life prediction result.

4. The environment difference-based lubricating oil aging prediction method according to claim 3, characterized by, The method of inputting the to-be-used environmental condition and the predicted use performance parameter into the lubricating oil remaining use life predictor to perform lubricating oil remaining use life prediction comprises: performing time series on the to-be-used environmental condition and the predicted use performance parameter to obtain to-be-used environmental condition time series data and predicted use performance parameter time series data; randomly extracting first to-be-used environmental variable time series data based on the to-be-used environmental condition time series data; randomly extracting first predicted use performance parameter time series data based on the predicted use performance parameter time series data; interacting the first to-be-used environmental variable time series data and the first predicted use performance parameter time series data to obtain random interaction time series data; performing time trend extraction on the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data, and performing lubricating oil remaining use life prediction based on the trend.

5. The environment difference-based lubricating oil aging prediction method according to claim 4, characterized by, The method of performing time trend extraction on the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data, and performing lubricating oil remaining use life prediction based on the trend comprises: extracting trend components, seasonal components and random components in the to-be-used environmental condition time series data, the predicted use performance parameter time series data and the random interaction time series data; performing characteristic capture on the trend components, the seasonal components and the random components respectively to obtain trend component characteristics, seasonal component characteristics and random component characteristics; performing lubricating oil remaining use life prediction based on the trend component characteristics, the seasonal component characteristics and the random component characteristics combined with weight distribution.

6. The environment difference-based lubricating oil aging prediction method according to claim 1, characterized by, The method of inputting the to-be-used environmental condition into the lubricating oil environmental differential aging prediction model to perform lubricating oil aging prediction, generating and outputting a lubricating oil aging prediction result further comprises: performing data enhancement on the lubricating oil use data set to obtain an enhanced lubricating oil use data set; testing the lubricating oil environmental differential aging prediction model using the enhanced lubricating oil use data set; correcting the lubricating oil environmental differential aging prediction model according to the test result.

Images