CN-121456852-B - Filter element life prediction method, electronic equipment and medium

CN121456852BCN 121456852 BCN121456852 BCN 121456852BCN-121456852-B

Abstract

The invention relates to the technical field of surface miner, in particular to a filter element service life prediction method, electronic equipment and a medium. According to the method, state information of the hydraulic circuit of the engineering machinery in a first preset time period is constructed into a plurality of sample features (each sample feature represents an operation state in a single preset unit time period), an original model is trained to obtain a life prediction model, so that the model can learn degradation rules and time sequence features of the filter element in different working condition states from a historical operation sample, real-time state information in a second preset time period is acquired again in actual application, the life prediction model is input, and the residual service life of the filter element after the second preset time period is directly output.

Inventors

ZHANG HAO
ZHANG YANHONG
QU HAILONG

Assignees

三一重型装备有限公司
三一智能装备有限公司

Dates

Publication Date: 20260505
Application Date: 20260106

Claims (9)

1. A method of predicting the life of a filter element, wherein the filter element is located in an oil return line of a hydraulic circuit of a construction machine, the method comprising: Acquiring state information of the hydraulic circuit in a first preset time period, wherein the state information in the first preset time period is used for representing the running state of the hydraulic circuit in the first preset time period; constructing a plurality of sample features by using state information in the first preset time period, wherein each sample feature is used for representing the running state of the hydraulic circuit in a single preset unit time period; performing model training on the original model by utilizing the plurality of sample characteristics to obtain a life prediction model; Acquiring state information of the hydraulic circuit in a second preset time period, and inputting the state information in the second preset time period into the life prediction model to obtain the residual service life of the filter element output by the life prediction model after the second preset time period, wherein the state information in the second preset time period is used for representing the running state of the hydraulic circuit in the second preset time period; The life prediction model comprises a working condition classification sub-model, a load spectrum prediction sub-model and a trunk prediction sub-model, wherein the model output of the working condition classification sub-model comprises working condition categories of the engineering machinery at each time point in the second preset time period, the model output of the load spectrum prediction sub-model comprises load spectrums of the engineering machinery at a plurality of time points after the second preset time period, and the trunk prediction sub-model is used for determining the residual service life of the filter element after the second preset time period according to state information in the second preset time period, model output of the working condition classification sub-model and model output of the load spectrum prediction sub-model.
2. The method of claim 1, wherein the status information for the first preset time period includes a pressure differential between an upstream pressure and a downstream pressure of the filter element for the first preset time period and an oil temperature of the oil return line for the first preset time period; the constructing a plurality of sample features by using the state information in the first preset time period includes: Preprocessing the state information in the first preset time period; Dividing the state information in the first preset time period after the pretreatment according to the preset unit time length to obtain the plurality of sample characteristics; Wherein the preprocessing comprises: For each time point in the first preset time period, determining the oil viscosity of the time point according to the oil temperature of the time point; and normalizing the pressure difference at the time point according to the oil viscosity.
3. The method of claim 1, wherein the remaining useful life is characterized by remaining life information comprising a remaining life specific value and a weibull distribution parameter with the remaining life specific value; The model training comprises a plurality of iteration cycles, and the model corresponding to each iteration cycle is a target model of the iteration cycle; the training the original model by using the plurality of sample features comprises the following steps: determining a loss value of each iteration period by utilizing a target loss function, and updating and iterating model parameters of a target model of the iteration period according to the loss value; The loss value of the target loss function is the sum value of a first loss value, a second loss value, a third loss value and a fourth loss value, the first loss value is the product of a first weight and the loss value of the first loss function, the first loss function is used for determining the loss value of a first original sub-model in each iteration period, and the first original sub-model corresponds to the working condition classification sub-model; the second loss value is the product of a second weight and a loss value of a second loss function, the second loss function is used for determining the loss value of a second original sub-model in each iteration period, and the second original sub-model corresponds to the load spectrum prediction sub-model; The third loss value is the product of a third weight and a loss value of a third loss function, the third loss function is used for determining a loss value of a third primary sub-model corresponding to the residual life specific value in each iteration period, and the third primary sub-model corresponds to the trunk prediction sub-model; The fourth loss value is the product of a fourth weight and a loss value of a fourth loss function, and the fourth loss function is used for determining the loss value of a third primitive sub-model corresponding to the Weibull distribution parameter of the specific residual life value in each iteration period.
4. A method according to claim 3, wherein the first loss function is based on cross entropy loss, the second loss function is based on quantile loss or mean square error loss, the third loss function is based on Hu Ba loss, and the fourth loss function comprises a first sub-loss function and a second sub-loss function; The first sub-loss function is used for determining a loss value of a Weibull distribution parameter corresponding to a first sample, the second sub-loss function is used for determining a loss value of a Weibull distribution parameter corresponding to a second sample, the first sub-loss function is a negative log likelihood of a probability density function of the Weibull distribution, and the second sub-loss function is a negative log likelihood of a survival function of the Weibull distribution; The first sample is a sample characteristic corresponding to a preset unit duration of which the filter element is invalid in the plurality of sample characteristics; The second sample is a sample characteristic corresponding to a preset unit duration in which the filter element is not invalid in the plurality of sample characteristics.
5. The method of claim 1, wherein the remaining useful life is characterized by remaining life information comprising a remaining life specific value and a weibull distribution parameter with the remaining life specific value; After the remaining service life of the filter element after the second preset period of time, from which the lifetime prediction model output is obtained, the method further includes: And determining residual life predicted values corresponding to a plurality of preset quantiles according to the Weibull distribution parameters of the residual life specific values.
6. The method of claim 5, wherein after determining the remaining life prediction values corresponding to the plurality of preset fractional numbers, the method further comprises: Determining a residual life predicted value corresponding to a first quantile and a residual life predicted value corresponding to a second quantile from residual life predicted values corresponding to the preset quantiles, wherein the first quantile is larger than the second quantile; and determining a recommended replacement time period according to the residual life predicted value corresponding to the first quantile, the residual life predicted value corresponding to the second quantile and the current time point, wherein the recommended replacement time period is used for indicating the time period for replacing the filter element.
7. The method of claim 1, wherein the remaining useful life is characterized by remaining life information comprising a remaining life specific value and a weibull distribution parameter with the remaining life specific value; After the remaining service life of the filter element after the second preset period of time, from which the lifetime prediction model output is obtained, the method further includes: Acquiring a pressure difference between upstream pressure and downstream pressure of the filter element, and determining the acquired pressure difference between the upstream pressure and the downstream pressure of the filter element as a current pressure difference; determining whether the probability that the residual life of the filter element is smaller than or equal to a preset life threshold is larger than or equal to a preset probability threshold according to the Weibull distribution parameter of the residual life specific value; outputting a filter element replacement prompt when the probability that the residual life of the filter element is smaller than or equal to the preset life threshold is larger than or equal to the preset probability threshold, or the current pressure difference meets a first preset condition, or the current pressure difference meets a second preset condition; The first preset condition is that the current pressure difference is larger than or equal to a first difference value, and the first difference value is a difference value obtained by subtracting a preset buffer value from a preset pressure difference threshold value; the second preset condition is that the current pressure difference is greater than or equal to an opening threshold value of a bypass valve of the oil return pipeline.
8. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus; a memory for storing a computer program; A processor for implementing the steps of the cartridge life prediction method of any one of claims 1-7 when executing a program stored on a memory.
9. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the cartridge life prediction method of any one of claims 1-7.

Description

Filter element life prediction method, electronic equipment and medium Technical Field The invention relates to the technical field of surface miner, in particular to a filter element service life prediction method, electronic equipment and a medium. Background Among various engineering machines, many engineering machines (such as surface mining equipment) operate under high dust, strong wind and sand, large day and night temperature difference and other severe environments for a long time under high load, and a filter element (namely an oil return filter element) in an oil return pipeline of a hydraulic system is easy to block due to the fact that the filter element is used for continuously intercepting particles and moisture, so that the service life of the engineering machines is prolonged. For the replacement of the filter element, a method of 'regular replacement' or 'single differential pressure threshold alarming to prompt replacement' is generally adopted at present, and complicated working sites, various working condition changes and other factors are not considered, so that the filter element is easy to replace too early or too late. Therefore, how to intelligently, scientifically and accurately estimate the residual life of the filter element is a key for helping technicians to determine the replacement time of the filter element. Disclosure of Invention The embodiment of the invention provides a filter element life prediction method, electronic equipment and a medium, and aims to intelligently, scientifically and accurately predict the residual life of a filter element. In a first aspect, an embodiment of the present application provides a method for predicting a life of a filter element, where the filter element is located in an oil return line of a hydraulic circuit of an engineering machine, the method including: Acquiring state information of the hydraulic circuit in a first preset time period, wherein the state information in the first preset time period is used for representing the running state of the hydraulic circuit in the first preset time period; constructing a plurality of sample features by using state information in the first preset time period, wherein each sample feature is used for representing the running state of the hydraulic circuit in a single preset unit time period; performing model training on the original model by utilizing the plurality of sample characteristics to obtain a life prediction model; Acquiring state information of the hydraulic circuit in a second preset time period, and inputting the state information in the second preset time period into the life prediction model to obtain the residual service life of the filter element output by the life prediction model after the second preset time period, wherein the state information in the second preset time period is used for representing the running state of the hydraulic circuit in the second preset time period. In some embodiments, the status information for the first preset time period includes a pressure difference between an upstream pressure and a downstream pressure of the filter element for the first preset time period and an oil temperature of the oil return line for the first preset time period; the constructing a plurality of sample features by using the state information in the first preset time period includes: Preprocessing the state information in the first preset time period; Dividing the state information in the first preset time period after the pretreatment according to the preset unit time length to obtain the plurality of sample characteristics; Wherein the preprocessing comprises: For each time point in the first preset time period, determining the oil viscosity of the time point according to the oil temperature of the time point; and normalizing the pressure difference at the time point according to the oil viscosity. In some embodiments, the life prediction model includes a working condition classification sub-model, a load spectrum prediction sub-model, and a trunk prediction sub-model, wherein the model output of the working condition classification sub-model includes a working condition category of the engineering machine at each time point within the second preset time period, the model output of the load spectrum prediction sub-model includes load spectrums of the engineering machine at a plurality of time points after the second preset time period, and the trunk prediction sub-model is used for determining the residual service life of the filter element after the second preset time period according to state information in the second preset time period, model output of the working condition classification sub-model, and model output of the load spectrum prediction sub-model. In some embodiments, the remaining useful life is characterized by remaining life information comprising a remaining life specific value and a weibull distribution parameter with the remaining life specific value;