CN114881321A - Mechanical component failure prediction method, device, electronic device and storage medium - Google Patents
Mechanical component failure prediction method, device, electronic device and storage medium Download PDFInfo
- Publication number
- CN114881321A CN114881321A CN202210479278.6A CN202210479278A CN114881321A CN 114881321 A CN114881321 A CN 114881321A CN 202210479278 A CN202210479278 A CN 202210479278A CN 114881321 A CN114881321 A CN 114881321A
- Authority
- CN
- China
- Prior art keywords
- time period
- sample
- condition information
- working condition
- fault
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000003860 storage Methods 0.000 title claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 61
- 238000010801 machine learning Methods 0.000 claims abstract description 22
- 238000000605 extraction Methods 0.000 claims description 54
- 230000008569 process Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 description 33
- 238000011156 evaluation Methods 0.000 description 22
- 238000004891 communication Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000001186 cumulative effect Effects 0.000 description 3
- 238000003066 decision tree Methods 0.000 description 3
- 206010016256 fatigue Diseases 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007858 starting material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Educational Administration (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Manufacturing & Machinery (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Software Systems (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The application relates to the technical field related to fault prediction, in particular to a mechanical component fault prediction method and device, electronic equipment and a storage medium. The mechanical component fault prediction method comprises the following steps: acquiring historical working condition information, wherein the historical working condition information comprises: working condition information from when the mechanical component is started to be used until the current moment; extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristics are used for characterizing the use condition of the mechanical part; inputting the use characteristics into a fault prediction model to obtain a prediction result; the fault prediction model is obtained by training a machine learning model which is built in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part. Whether the mechanical part fails within the preset time period from the current moment is predicted through the failure prediction model.
Description
Technical Field
The application relates to the technical field related to fault prediction, in particular to a mechanical component fault prediction method and device, electronic equipment and a storage medium.
Background
With the progress of science and technology, more and more machines enter the production and life of people. In the use of these machines, some of the mechanical components of the machines may fail. The use of these machines by people is greatly affected when these machine components suddenly fail. For example: if the mechanical parts of the crane suddenly break down, the crane cannot be used normally, and a series of production plans related to the crane are influenced.
However, there is currently no practical means to predict whether these mechanical components will fail. There is therefore an urgent need for a means to predict whether a mechanical component will fail.
Disclosure of Invention
In view of the above, embodiments of the present application are directed to a method, an apparatus, an electronic device, and a storage medium for predicting a failure of a mechanical component in a future period of time.
According to a first aspect of embodiments of the present application, there is provided a method for predicting a failure of a mechanical component, including:
acquiring historical working condition information, wherein the historical working condition information comprises: operating condition information from when the mechanical component is started to be used until the current time;
extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristic is used for characterizing the use condition of the mechanical part;
obtaining a prediction result based on the use characteristics and the fault prediction model;
the fault prediction model is obtained by training a machine learning model which is built in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
In one embodiment, the training process of the fault prediction model includes:
obtaining a first preset number of sample use characteristics and a sample identification corresponding to each sample use characteristic as training samples;
the sample identification is used for characterizing a mechanical part corresponding to the sample use characteristic, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical working condition information of the sample use characteristics is extracted;
and training a machine learning model which is set up in advance based on the training sample to obtain a fault prediction model.
In one embodiment, the obtaining of the first preset number of sample usage characteristics and the sample identification corresponding to each sample usage characteristic as training samples includes:
acquiring multiple groups of sample historical working condition information and information about whether a mechanical part corresponding to each group of sample historical working condition information has a fault;
each group of sample historical working condition information comprises data which is generated in the using process of a mechanical part and used for representing the working condition information;
determining data related to whether the mechanical part fails in the sample historical working condition information as target data;
extracting sample use characteristics based on the target data for each group of the historical working condition information;
and determining a sample identification corresponding to the sample use characteristic for each sample use characteristic.
In one embodiment, the determining data related to whether the mechanical component is faulty in the sample historical operating condition information is target data, and includes:
obtaining relevance information of sample historical working condition information, and determining whether data related to the fault of the mechanical component in the sample historical working condition information is target data based on the relevance information, wherein the relevance information comprises information used for indicating whether the data related to the fault of the mechanical component in the sample historical working condition information input by related personnel; or the like, or, alternatively,
and calculating a correlation coefficient between the data in the sample historical working condition information and whether the mechanical part fails, and determining the data related to whether the mechanical part fails in the sample historical working condition information as target data based on the correlation coefficient.
In one embodiment, said extracting sample usage characteristics based on said target data comprises:
target data corresponding to each basic time period is summarized to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
determining a characteristic extraction time period which starts from the time when the mechanical part starts to be used and ends from the end time of the basic time period for each basic time period;
and determining a basic time period contained in each feature extraction time period as a target basic time period, and performing feature extraction and summarization on the corresponding working condition information of the target basic time period to obtain the use features of the sample.
In one embodiment, further comprising: evaluating the fault prediction model to obtain an evaluation index;
judging whether the evaluation index reaches a preset index or not;
and if the evaluation index does not reach a preset index, retraining the fault prediction model.
In one embodiment, the extracting the use characteristics of the mechanical component based on the historical operating condition information includes:
summarizing historical working condition information in the basic time period aiming at each basic time period to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
and summarizing the corresponding working condition information of each basic time period, and extracting the characteristics to obtain the use characteristics.
According to a second aspect of embodiments of the present application, there is provided a mechanical component failure prediction apparatus including:
the acquisition module is used for acquiring historical working condition information, and the historical working condition information comprises: operating condition information from when the mechanical component is started to be used until the current time;
the extraction module is used for extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristic is used for characterizing the use condition of the mechanical part;
the input module is used for obtaining a prediction result based on the use characteristic and the fault prediction model;
the fault prediction model is obtained by training a machine learning model which is built in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
In one embodiment, the mechanical component failure prediction apparatus further comprises: a training module for training the fault prediction model;
the training module is specifically configured to: obtaining a first preset number of sample use characteristics and a sample identification corresponding to each sample use characteristic as training samples; the sample identification is used for characterizing a mechanical part corresponding to the sample use characteristic, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical working condition information of the sample use characteristics is extracted; and training the machine learning model based on the training samples to obtain the fault prediction model.
According to a third aspect of embodiments herein, there is provided an electronic device comprising:
a processor; a memory for storing the processor-executable instructions; the processor is configured to perform the method according to any of the above embodiments.
According to a fourth aspect of embodiments herein, there is provided a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to perform a method as in any one of the above embodiments.
In the method for predicting the fault of the mechanical component provided by the embodiment of the application, historical working condition information is firstly acquired, and the historical working condition information comprises the following steps: the method comprises the steps that working condition information from the beginning of use of a mechanical part to the current moment can be obtained, wherein historical working condition information can reflect the accumulated use condition of the mechanical part, then the use characteristics of the mechanical part are extracted based on the historical working condition information, and the use characteristics are input into a pre-trained fault prediction model to obtain a prediction result; the prediction result is used for indicating whether the mechanical part fails within a preset time period from the current moment. By means of the arrangement, whether the mechanical part fails within a future period of time can be predicted by means of the failure prediction model based on all actual historical use conditions of the mechanical part.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.
Fig. 1 is a schematic flow chart illustrating a method for predicting a failure of a mechanical component according to an embodiment of the present application.
Fig. 2 is a schematic flow chart illustrating a method for predicting a failure of a mechanical component according to an embodiment of the present application.
Fig. 3 is a partial schematic flow chart of a method for predicting a failure of a mechanical component according to an embodiment of the present application.
Fig. 4 is a schematic time division diagram of a method for predicting a failure of a mechanical component according to an embodiment of the present application.
Fig. 5 is a partial schematic flow chart of a method for predicting a failure of a mechanical component according to an embodiment of the present application.
Fig. 6 is a block diagram illustrating a mechanical component failure prediction apparatus according to an embodiment of the present application.
Fig. 7 is a block diagram illustrating an electronic device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Summary of the application
With the progress of science and technology, more and more machines enter the production and life of people. In the use of these machines, some of the mechanical components of the machines may fail. The use of these machines by people is greatly affected when these machine components suddenly fail. For example: if the mechanical parts of the crane suddenly break down, the crane cannot be used normally, and a series of production plans related to the crane are influenced. The mechanical component of the crane can be, but is not limited to, a power take-off transmission shaft. However, there is currently no practical means to predict whether these mechanical components will fail. There is therefore an urgent need for a means to predict whether a mechanical component will fail.
In order to solve the above problems, in the embodiments of the present application, based on whether a mechanical component of an apparatus fails or not and a specific use condition of the apparatus have strongly related characteristics, a large amount of historical operating condition data in use of the apparatus is collected, so as to characterize a specific cumulative use condition of the mechanical component of the apparatus based on the historical operating condition data of the apparatus, and based on the cumulative use condition of the apparatus, a pre-trained model is used to predict whether the mechanical component will fail or not within a period of time in the future.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary method
Fig. 1 is a schematic flow chart of a method for predicting a failure of a mechanical component according to an embodiment of the present disclosure, and as shown in fig. 1, the method includes the following steps.
S110, obtaining historical working condition information, wherein the historical working condition information comprises: the operating condition information from when the mechanical component is started to be used until the present time.
It should be noted that the condition information refers to all information related to the machine that can be acquired and determined. For a device, the operating condition information of the device comprises: various information collected by the device and various instructions for the device. The use of the machine refers to various conditions that affect the service life of the machine. When a mechanical component is used, the mechanical component is generally mounted on equipment, and in this case, the historical operating condition information of the equipment can be regarded as the historical operating condition information of the mechanical component. For example: the power take-off transmission shaft is used as a mechanical component and is generally installed on a crane. In the use of the crane, the power take-off transmission shaft plays a role in force transmission. The historical working condition information of the crane can be regarded as the historical working condition information of the power take-off transmission shaft. Furthermore, the equipment generally has the function of collecting the historical working condition information of the equipment, so the historical working condition information collected by the equipment can be directly obtained. Still taking the power take-off transmission shaft installed on the crane as an example for explanation, the working condition information from the beginning of the use of the mechanical component to the current time includes: ECU data related to the power take-off transmission shaft are collected by a vehicle-mounted sensor T-BOX of the crane from the moment the power take-off transmission shaft is installed on the crane to the current moment.
S120, extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristics are used for characterizing the use condition of the mechanical part;
it should be noted that there is a great amount of useless information and discrete information in the acquired historical operating condition information. Therefore, the historical operating condition information is directly used for prediction, not only the data amount is overlarge, but also the machine learning model is difficult to extract knowledge related to whether the mechanical part fails or not due to the fact that the data are discrete, so that the problems of low prediction efficiency, low prediction result accuracy and the like are caused.
And S130, inputting the use characteristics into a fault prediction model to obtain a prediction result.
The fault prediction model is obtained by training a machine learning model which is built in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
Through the steps, whether the mechanical part fails within a period of time in the future or not can be predicted by means of the failure prediction model based on all actual historical use conditions of the mechanical part, and convenience is brought to use planning of equipment.
In the solution provided in the present application, prediction needs to be performed by using a fault prediction model obtained through pre-training. The training effect of the fault prediction model determines the accuracy of the prediction result when prediction is carried out. Referring to fig. 2, the training process of the fault prediction model may include:
s210, obtaining a first preset number of sample use characteristics and sample identifications corresponding to each sample use characteristic as training samples.
The sample identification is used for representing a mechanical part corresponding to the use characteristic of the sample, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical operating condition information of the sample use characteristics is extracted.
For example: during the use process of a crane, the power take-off transmission shaft can be damaged, and the working condition information of the crane in the time period from the time when the power take-off transmission shaft is installed on the crane and used to the time when the power take-off transmission shaft is in failure can be selected so as to reflect the use condition of the power take-off transmission shaft based on the information. It should be noted that, based on different devices corresponding to mechanical components, the extracted operating condition information is different, and the following description is given by taking a power take-off transmission shaft mounted on a crane as an example: the historical operating condition information comprises a large amount of ECU data collected by crane vehicle-mounted sensors. Specifically, the historical operating condition information includes: and the working time of getting on the bus, the rotating speed of the engine, the output torque of the engine, the power taking signal of getting on the bus and the like. It should be noted that historical operating condition data generated in the using process of the crane with no failure of the power take-off transmission shaft can also be collected to generate corresponding using characteristics.
And S220, training the pre-built machine learning model based on the training sample to obtain a fault prediction model.
It should be noted that the pre-built machine learning model may be a LightGBM model. Of course, the machine learning model may be other machine learning models. The core of the machine learning model in the embodiment of the application is to receive training and then complete prediction. Thus, the pre-built machine learning model may be built based on other machine learning algorithms or a combination of multiple algorithm models.
Furthermore, in order to enable the trained fault prediction model to have higher prediction accuracy, the fault prediction model can be evaluated, and if the fault prediction model is poor in prediction effect, the fault prediction model can be retrained. Specifically, referring to fig. 2, the correlation process of the evaluation is as follows:
and S230, evaluating the fault prediction model to obtain an evaluation index.
When the failure prediction model is evaluated, the simulation is predicted by the failure prediction model. And constructing use characteristic set data which is distributed similarly to the use characteristic set data in practical application, wherein the part of the use characteristic set data has corresponding prediction result identification, inputting the part of the use characteristic set data into a fault prediction model to obtain a prediction result, comparing the prediction result with the prediction result identification, and summarizing the comparison result to obtain an evaluation index for representing whether the preset data is accurate or not.
S240, judging whether the evaluation index reaches a preset index or not.
And S250, if the evaluation index does not reach the preset index, retraining the fault prediction model.
If the evaluation index does not reach the preset index, the current fault prediction model cannot achieve the expected prediction effect, and the fault prediction model needs to be retrained until the trained fault prediction model has higher prediction accuracy.
Based on the steps, the training of the fault prediction model can be completed, the trained fault prediction model is evaluated, and if the training effect of the fault prediction model is not ideal, the fault prediction model is retrained, so that the fault prediction model can have high prediction accuracy when actual prediction is carried out.
Specifically, since the historical operating condition information of the mechanical component is generally complex, it is a complex process to "obtain the first preset number of sample usage characteristics and the sample identification corresponding to each sample usage characteristic as the training sample". Specifically, referring to fig. 3, step S210 may include:
s211, acquiring multiple groups of sample historical working condition information and information whether the mechanical part corresponding to each group of sample historical working condition information has a fault.
Specifically, historical operating condition information of the mechanical component is obtained and used as sample historical operating condition information. The historical operating condition information corresponding to one mechanical part is a group of sample historical operating condition information: taking a power take-off transmission shaft of a crane as an example for explanation: firstly, extracting a fault equipment list of power take-off transmission shaft fracture faults caused by high load and untimely maintenance according to after-sales fault information of a user, intercepting historical ECU data of fault equipment from delivery time or last replacement time to fault occurrence time as sample historical working condition information from a large amount of ECU data (such as vehicle-mounted working time, engine rotating speed, engine output torque, vehicle-mounted power take-off signals and the like) collected by corresponding crane vehicle-mounted sensors. Because the equipment state when the fault occurs is the fault state, if the predictable cycle is defined as one month, the ECU data from the factory time or the last maintenance time to the month before the fault occurs is intercepted from the historical ECU data and is used as the data basis for predicting whether the equipment has the fault in the next month. And characterizing the current use state of the equipment by extracting historical statistical characteristics of the working condition of the vehicle-on operation from ECU data of one month before the fault occurs.
S212, determining whether data related to the fault of the mechanical part in the sample historical working condition information is target data;
in practical application, because the sample historical operating condition information contains a large amount of data of various types, wherein a part of the data is irrelevant or basically irrelevant to whether the mechanical part has a fault, the part of the data can be eliminated, and only the data relevant to whether the mechanical part has the fault is used as target data to participate in the subsequent process.
Taking a power take-off transmission shaft as an example: the regional statistics of the working condition intervals are carried out according to the rotating speed of the engine and the output torque of the engine, and the crane boarding operation can be divided into a common working condition interval (the rotating speed is 600-900 rpm, and the torque is 0-800N.m) and a large rotating speed and high torque working condition interval (the rotating speed is greater than 900rpm or the torque is greater than 800N.m) according to the rotating speed and the torque range. According to the historical rotating speed and torque thermodynamic summary result of the power take-off transmission shaft fault vehicle, the fault vehicle is more prone to having a larger working time ratio in a large rotating speed and high torque interval, and the summary result shows that the rotating speed and torque range is data related to whether a mechanical part is in fault or not. Therefore, when the target data is determined, the inter-partition statistical data according to the rotational speed and the torque may be used as a part of the target data.
Specifically, the specific manner of determining whether the data related to the failure of the mechanical component in the sample historical operating condition information is the target data may include the following two manners:
one of the methods is as follows: and acquiring the correlation information of the sample historical working condition information, and determining whether the data related to the fault of the mechanical component in the sample historical working condition information is target data based on the correlation information, wherein the correlation information comprises information used for indicating whether the data related to the fault of the mechanical component in the sample historical working condition information input by related personnel is faulty. In the above manner, involvement of the relevant staff is required, and the relevant staff determines whether the data is relevant to the failure of the mechanical component and whether the data should be used to make a prediction of the mechanical component. In this way, the determination of the target data can be completed quickly by the experience of the relevant staff.
Taking the power take-off transmission shaft as an example, the fatigue failure mechanism analysis can be carried out according to the fatigue failure mechanism analysis of the power take-off transmission shaft, under the working conditions of the power take-off and pressure test block of the oil pump, the instantaneous impact load and the continuous high load generated when the hook is lifted, the hook is fallen and the accelerator is instantly increased and reduced, the power take-off transmission shaft is easy to generate early fatigue and distortion deformation, and the temperature of a cross joint is increased, so that the ten-byte abrasion and fracture of the power take-off transmission shaft are aggravated. Therefore, the frequency of the instantaneous impact working condition, the working time length of getting on the vehicle under the working condition interval with high rotating speed and large rotating speed of the engine, the statistical mean value of the torque in the interval, the mean value of the power of the engine during the working of getting on the vehicle, the median, the accumulated value and the like are all related to whether the power taking transmission shaft fails, and based on the data, related workers can indicate the data related to whether the power taking transmission shaft fails or not in a mode of inputting the correlation degree information, so that the characteristic extraction statistical characteristics are carried out on the corresponding working condition information of the target basic time period through the working conditions with different dimensions to describe the load state of the power taking transmission shaft.
The other mode is as follows: and calculating a correlation coefficient between the data in the sample historical working condition information and whether the mechanical part fails, and determining the data related to whether the mechanical part fails in the sample historical working condition information as target data based on the correlation coefficient. The above manner is to reflect whether or not the data is related to the failure of the mechanical component by calculating the correlation coefficient. In this way, the influence of human factors on the determination result can be avoided when the target data is determined, and the target data is determined through the existing strategy for calculating the correlation coefficient, so that the determined target data is more objective and accurate.
S213, extracting the use characteristics of the sample according to each group of historical working condition information and based on target data;
it should be noted that, for each group of historical operating condition information, based on the target data, a plurality of groups of sample use characteristics can be extracted;
in a particular application, extracting sample usage characteristics based on the target data may include:
firstly, target data corresponding to each basic time period is summarized to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length; in addition, in the target data, a part of the data is sufficiently discrete, and in order to collect the discrete part of the data, the use time period of the mechanical component may be preferentially divided into segments of a preset time length which are connected end to end, so that a plurality of basic time periods are obtained. And then summarizing the discrete data in each basic time period group, and then summarizing the data based on the basic time periods so as to summarize the discrete data.
Taking a power take-off transmission shaft of a crane as an example: due to the fact that the crane has the operation characteristics of irregularity, discontinuity and the like, the time scale of statistics of the target data is not too small, and the time scale of the preset time period is predicted to be not too small in the same way. The statistical time scale for the target data, i.e. the length of the base time period, may be 10 days, 15 days, 20 days or 30 days. The length of the preset time period may be 20 days or 30 days. It should be noted that the time length of the preset time period needs to be greater than or equal to the length of the basic time period.
Specifically, with reference to fig. 4, the total time corresponding to the target data is the time from "the time when the mechanical component starts to be used" to "the failure time", and this part of time is divided into 7 basic time periods and part of remaining time (for ease of understanding, this part of remaining time is complemented to basic time period 8 in fig. 4).
Then, for each basic time period, determining a characteristic extraction time period which starts from the time when the mechanical part starts to be used and ends from the end time of the basic time period;
the description is made with the example in fig. 4: the feature extraction period includes: a feature extraction period constituted by "basic period 1"; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 2"; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 3"; a feature extraction time period consisting of "the starting time of the basic time period 1" to "the end time of the basic time period 4"; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 5"; a feature extraction time period consisting of "the starting time of the basic time period 1" to "the end time of the basic time period 6"; and a feature extraction period consisting of "the start time of the basic period 1" to "the end time of the basic period 7". According to the arrangement, a plurality of feature extraction time periods can be determined based on the same group of historical working condition data, and then one sample use feature is extracted based on each feature extraction time period. It should be noted that, in order to reduce the training data, only part of the feature extraction period may be abandoned, for example: a feature extraction period constituted by "basic period 1"; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 2"; and a feature extraction period consisting of "the start time of the basic period 1" to "the end time of the basic period 3". The part of the feature extraction time period is short, and the extracted use features have no great guidance effect on predicting whether the mechanical part is in failure or not, so that the part of the feature extraction time period can be abandoned.
And then, determining a basic time period contained in the feature extraction time period as a target basic time period for each feature extraction time period, and performing feature extraction and summarization on working condition information corresponding to the target basic time period to obtain the sample use features.
The description is made with the example in fig. 4: the feature extraction period constituted by "the start time of the basic period 1" to "the end time of the basic period 6" includes target basic periods: a basic period 1, a basic period 2, a basic period 3, a basic period 4, a basic period 5, and a basic period 6.
Taking a power take-off transmission shaft of a crane as an example for explanation, the working time length ratio, the average value and the standard deviation of the output torque of the engine, the average value and the standard deviation of the rotating speed of each basic time period of each vehicle in each interval are respectively calculated. Meanwhile, the power sum, the power mean and the power standard deviation of each vehicle in each basic time period are calculated. Then, the characteristics calculated for each vehicle score basic time segment are combined: build the cumulative signature over the entire life cycle: the accumulated value of the frequency of the instantaneous impact working condition, the accumulated working time of the vehicle getting on and the power sum. Meanwhile, an accumulated value of the frequency of the instantaneous impact working condition is calculated, and the accumulated characteristic is obtained by the ratio of the power sum to the accumulated working time of getting on the vehicle.
S214, determining a sample identifier corresponding to the sample using characteristics aiming at each sample using characteristic.
Still taking fig. 4 as an example for explanation, based on fig. 4, it can be determined that no fault occurs in the preset time period after the feature extraction time period formed by "the starting time of the basic time period 1" to "the end time of the basic time period 2", and therefore the sample identifier corresponding to the usage feature extracted in the feature extraction time period is used to indicate that no fault occurs in the preset time period in the future; after a feature extraction time period formed by the starting time of the basic time period 1 to the end time of the basic time period 3, no fault occurs in a preset time period, so that a sample identifier corresponding to the use feature extracted in the feature extraction time period is used for indicating that no fault occurs in the future preset time period; the method comprises the steps that no fault occurs in a preset time period after a feature extraction time period formed by 'starting time of basic time period 1' to 'end time of basic time period 4', and therefore a sample identifier corresponding to a use feature extracted in the feature extraction time period is used for indicating that no fault occurs in the future preset time period; the method comprises the steps that no fault occurs in a preset time period after a feature extraction time period formed by 'starting time of basic time period 1' to 'end time of basic time period 5', and therefore a sample identifier corresponding to a use feature extracted in the feature extraction time period is used for indicating that no fault occurs in the future preset time period; a fault occurs in a preset time period after a feature extraction time period formed by a starting time of a basic time period 1 to a terminal time of the basic time period 6, so that a sample identifier corresponding to a use feature extracted in the feature extraction time period is used for indicating that the fault occurs in the future preset time period; a fault occurs in a preset time period after the feature extraction time period formed by the "starting time of the basic time period 1" to the "ending time of the basic time period 7", and therefore the sample identifier corresponding to the usage feature extracted in the feature extraction time period is used for indicating that a fault occurs in the future preset time period.
It should be noted that the usage characteristics may also be extracted based on mechanical components that have not failed all the time. Specifically, referring to fig. 4, if the time indicated by the failure time in fig. 4 is the current time rather than the failure time: a feature extraction time period consisting of "basic time period 1" may be selected; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 2"; a feature extraction time period formed from "the starting time of the basic time period 1" to "the end time of the basic time period 3"; a feature extraction time period consisting of "the starting time of the basic time period 1" to "the end time of the basic time period 4"; and extracting the use characteristics in a characteristic extraction time period from the starting time of the basic time period 1 to the end time of the basic time period 5, wherein the sample identification corresponding to the extracted use characteristics is used for indicating that no fault occurs in the future preset time period.
Extracting the statistical characteristics of the same type of equipment, and forming a fault characteristic matrix for the use characteristics of faults occurring in a future preset time period based on the corresponding identification
Wherein m is the number of characteristic parameters, and n is the number of samples of the fault equipment. Similarly, a normal feature matrix is formed for the use features which are not failed in the future preset time period based on the corresponding identification
Wherein m is characterizedThe number of parameters, k, is the number of samples of non-faulty devices.
When training of the model is actually performed, the feature matrices X and X may be combined 0 And inputting the corresponding prediction labels (fault or non-fault) into a LightGBM model for training, screening out sample characteristics which can most distinguish different prediction labels according to the characteristic matrix and the corresponding prediction labels by the decision tree model, calculating to obtain classification threshold values of the corresponding characteristics, and finally generating a decision tree which can classify the samples of the different prediction labels based on the current training sample data, wherein the decision tree can be used as a prediction rule for predicting whether the power take-off transmission shaft of the equipment has faults or not in a future period of time. Specifically, the model outputs a prediction probability for the fault label, so a probability threshold value can be set, and when the fault occurrence probability is greater than a discrimination threshold value, the prediction test sample is a fault vehicle, otherwise, the test sample is a non-fault vehicle. In order to optimize the final prediction effect, the probability threshold needs to be optimized by an optimization algorithm to obtain the optimal probability threshold.
The method for predicting the fault of the mechanical component provided by the application is further described by referring to fig. 5 in combination with a power take-off transmission shaft of a crane. It should be noted that the scheme shown in fig. 5 is mainly divided into two major parts, one is training and determining of a failure prediction model (see steps S501 to S506 in fig. 5 in detail), and the other is predicting whether a failure of a mechanical component (power take-off transmission shaft) occurs within a preset time period from the current time by using the determined failure prediction model (see steps S501 to S508 in fig. 5 in detail). Specifically, the method for predicting the fault of the mechanical component may include:
and S501, obtaining sample historical working condition data.
Specifically, historical working condition data of the power take-off transmission shafts are obtained in batches.
And S502, performing data cleaning on the historical working condition data.
It should be noted that the process of data cleansing includes: and removing obvious and unreasonable data in the historical working condition data, carrying out normalization processing on the data, determining target data and the like so as to be convenient for calling the historical working condition data in the following steps.
And S503, extracting the use characteristics and the sample identifications corresponding to the use characteristics as training data based on the historical working condition data.
It should be noted that, the specific process of "extracting the usage characteristic and the sample identifier corresponding to the usage characteristic based on the historical operating condition data" may refer to the related expression in the foregoing.
S504, training a pre-constructed machine learning model based on the training data to obtain a fault prediction model.
It should be noted that, the specific process of "training the pre-constructed machine learning model based on the training data to obtain the fault prediction model" may refer to the related expression in the foregoing.
And S505, calculating to obtain an evaluation index of the fault prediction model.
Specifically, the calculation process includes: determining accuracy, recall rate and comprehensive evaluation indexes through a confusion matrix; wherein the accuracy ratio Precision is TP/(TP + FP), the Recall ratio Recall is TP/(TP + FN), and the overall evaluation index F β is (1+ β) 2 )*Precision*Recall/(β 2 Precision) + Recall (β 1-5, where TP: true and true; FP: false positive example; FN: false counterexample; TN true negative) to evaluate the model results. The comprehensive evaluation index F beta is a weighted harmonic average value of the accuracy rate and the recall rate, and can comprehensively evaluate the model result by comprehensively balancing the accuracy rate and the recall rate. Therefore, the comprehensive evaluation index can be used as an evaluation index; because the fault prediction algorithm needs to predict as many faulty vehicles as possible under the condition of ensuring a certain accuracy, the model needs to have a higher recall rate, and the influence weight of the recall rate on the F beta index is higher by adjusting the weight beta value.
Wherein the confusion matrix is shown in the following table:
and S506, judging whether the evaluation index meets the requirement.
Namely: and determining whether the prediction result can be used or not by judging whether the comprehensive evaluation index meets the requirement, and when the prediction result cannot be used, re-executing the steps S502 to S506 until the comprehensive evaluation index meets the requirement, so that the prediction result of the fault prediction model is relatively fit with the actual situation at the moment, and the prediction result of whether the vehicle fails within the preset time has higher reliability.
Further, in the process of re-executing step S502 to step S506, the flow of cleaning the historical operating condition data may be modified, and the strategy for determining the relevant data may also be modified, so that the prediction result of the obtained fault prediction model is more accurate.
And S507, acquiring working condition data of the target to be predicted.
It should be noted that the strategy adopted when the condition data of the target to be predicted is obtained here should correspond to the strategy adopted when the model is trained.
And S508, extracting the use characteristics of the mechanical part based on the working condition data.
It should be noted that the strategy adopted when the condition data of the target to be predicted is obtained here should correspond to the strategy adopted when the model is trained.
Specifically, for each basic time period, summarizing historical working condition information in the basic time period to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length; and summarizing the corresponding working condition information of each basic time period, and extracting the characteristics to obtain the use characteristics. In this way, the use characteristics are extracted by adopting the strategy corresponding to the training model, so that the use characteristics corresponding to the training can be obtained, and the use characteristics for prediction are matched with the fault prediction model.
S509, inputting the use characteristics into a fault prediction model to obtain a prediction result;
based on the scheme, the prediction result of whether the mechanical part fails in a future period of time can be obtained, so that the equipment can be conveniently arranged to work based on the prediction result, and when the mechanical part of the equipment fails suddenly, the equipment can be timely processed without affecting the overall efficiency. Taking the power take-off transmission shaft of the crane as an example, if it is predicted that a part of starters will fail in the next month, other cranes can be allocated in advance to replace the crane when the crane fails, or the power take-off transmission shaft is prepared in advance so as to be convenient for timely maintenance of the crane.
Exemplary devices
The embodiment of the device can be used for executing the embodiment of the method. For details which are not disclosed in the embodiments of the apparatus of the present application, reference is made to the embodiments of the method of the present application.
Fig. 6 is a block diagram illustrating a mechanical component failure prediction apparatus according to an embodiment of the present application. As shown in fig. 6, the apparatus includes:
an obtaining module 601, configured to obtain historical operating condition information, where the historical operating condition information includes: working condition information from when the mechanical component is started to be used until the current moment;
an extraction module 602, configured to extract usage characteristics of the mechanical component based on historical operating condition information; the use characteristics are used for characterizing the use condition of the mechanical part;
an input module 603, configured to input the usage characteristics into the fault prediction model to obtain a prediction result;
the fault prediction model is obtained by training a machine learning model which is built in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
In one embodiment, the system further comprises a training module for completing the training of the fault prediction model; the training module is specifically configured to:
obtaining a first preset number of sample use characteristics and a sample identification corresponding to each sample use characteristic as training samples;
the sample identification is used for representing a mechanical part corresponding to the use characteristic of the sample, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical working condition information of the sample use characteristics is extracted;
and training a machine learning model which is set up in advance based on the training sample to obtain a fault prediction model.
In one embodiment, the training module is specifically configured to, when "the obtained first preset number of sample usage characteristics and the sample identifier corresponding to each sample usage characteristic are taken as training samples":
acquiring multiple groups of sample historical working condition information and information whether mechanical parts corresponding to each group of sample historical working condition information have faults or not;
each group of sample historical working condition information comprises data which is generated in the using process of one mechanical component and used for representing the working condition information;
determining data related to whether the mechanical part fails in the sample historical working condition information as target data;
extracting sample use characteristics based on target data according to each group of historical working condition information;
and determining a sample identification corresponding to the sample use characteristic for each sample use characteristic.
The training module is specifically used for executing the following steps when determining whether the data related to the fault of the mechanical component in the sample historical working condition information is the target data:
obtaining correlation information of sample historical working condition information, and determining whether data related to the fault of the mechanical component in the sample historical working condition information is target data based on the correlation information, wherein the correlation information comprises information used for indicating whether the data related to the fault of the mechanical component in the sample historical working condition information input by related personnel is faulty; or the like, or, alternatively,
and calculating a correlation coefficient between the data in the sample historical working condition information and whether the mechanical part fails, and determining the data related to whether the mechanical part fails in the sample historical working condition information as target data based on the correlation coefficient.
In one embodiment, the training module is specifically configured to, when performing "extracting the sample usage feature based on the target data":
target data corresponding to each basic time period is summarized to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
determining a characteristic extraction time period which starts from the time when the mechanical part starts to be used and ends from the end time of the basic time period for each basic time period;
and determining a basic time period contained in the feature extraction time period as a target basic time period for each feature extraction time period, and performing feature extraction and summarization on the corresponding working condition information of the target basic time period to obtain the use features of the sample.
In one embodiment, the fault prediction system further comprises an evaluation module, configured to evaluate the fault prediction model to obtain an evaluation index, determine whether the evaluation index reaches a preset index, and retrain the fault prediction model if the evaluation index does not reach the preset index.
In one embodiment, the extraction module is specifically configured to:
for each basic time period, summarizing historical working condition information in the basic time period to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
and summarizing the corresponding working condition information of each basic time period, and extracting the characteristics to obtain the use characteristics.
Exemplary electronic device
Referring to fig. 7, fig. 7 is a block diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 7, the electronic device may include: at least one processor 710, at least one communication interface 720, at least one memory 730, and at least one communication bus 740.
In the embodiment of the present invention, the number of the processor 710, the communication interface 720, the memory 730 and the communication bus 740 is at least one, and the processor 710, the communication interface 720 and the memory 730 are communicated with each other through the communication bus 740; it will be appreciated that the communication connections shown by processor 710, communication interface 720, memory 730, and communication bus 740 shown in FIG. 7 are merely optional.
The memory 730 stores application programs, may include a high-speed RAM memory, and may further include a non-volatile memory (non-volatile memory), such as at least one disk memory.
The processor 710 is specifically configured to execute an application program in the memory to implement any embodiment of the method for predicting a failure of a mechanical component described above.
Exemplary computer program product and computer-readable storage Medium
In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the method of mechanical component failure prediction according to various embodiments of the present application described in the "exemplary methods" section of this specification, supra.
The computer program product may be written with program code for performing the operations of embodiments of the present application in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, cause the processor to perform steps in a method of mechanical component failure prediction according to various embodiments of the present application described in the "exemplary methods" section above in this specification.
The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. A method of predicting a failure of a mechanical component, comprising:
acquiring historical working condition information, wherein the historical working condition information comprises: operating condition information from when the mechanical component is started to be used until the current time;
extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristic is used for characterizing the use condition of the mechanical part;
obtaining a prediction result based on the use characteristics and the fault prediction model;
the fault prediction model is obtained by training a pre-built machine learning model and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
2. The method of claim 1, wherein the training of the fault prediction model comprises:
obtaining a first preset number of sample use characteristics and a sample identification corresponding to each sample use characteristic as training samples;
the sample identification is used for characterizing a mechanical part corresponding to the sample use characteristic, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical working condition information of the sample use characteristics is extracted;
and training the machine learning model based on the training samples to obtain the fault prediction model.
3. The method of predicting a failure of a mechanical component according to claim 2, wherein said obtaining a first preset number of sample usage characteristics and a sample identifier corresponding to each sample usage characteristic as training samples comprises:
acquiring multiple groups of sample historical working condition information and information about whether a mechanical part corresponding to each group of sample historical working condition information has a fault;
each group of sample historical working condition information comprises data which is generated in the using process of a mechanical part and used for representing the working condition information;
determining data related to whether the mechanical part fails in the sample historical working condition information as target data;
extracting sample use characteristics based on the target data for each group of the historical working condition information;
and determining a sample identification corresponding to the sample use characteristic for each sample use characteristic.
4. The method of predicting a fault in a mechanical component of claim 3, wherein the determining whether the data related to the fault in the mechanical component in the sample historical operating condition information is target data comprises:
obtaining relevance information of the sample historical working condition information, and determining whether data related to the fault of the mechanical component in the sample historical working condition information is target data based on the relevance information, wherein the relevance information comprises information used for indicating whether the data related to the fault of the mechanical component in the sample historical working condition information input by related personnel; or the like, or, alternatively,
and calculating a correlation coefficient between the data in the sample historical working condition information and whether the mechanical part fails, and determining the data related to whether the mechanical part fails in the sample historical working condition information as target data based on the correlation coefficient.
5. The method of predicting a failure of a mechanical component of claim 3, wherein said extracting a sample usage signature based on said target data comprises:
target data corresponding to each basic time period is summarized to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
for each basic time segment, determining a characteristic extraction time segment which starts from the time when the mechanical part starts to be used and ends from the end time of the basic time segment;
and determining the basic time period contained in the feature extraction time period as a target basic time period aiming at each feature extraction time period, and performing feature extraction and summarization on the corresponding working condition information of the target basic time period to obtain the use features of the sample.
6. The method of predicting a fault in a mechanical component of claim 1, wherein said extracting a usage characteristic of the mechanical component based on the historical operating condition information comprises:
summarizing historical working condition information in the basic time period aiming at each basic time period to obtain working condition information corresponding to the basic time period; the basic time period is obtained by dividing the use time period of the mechanical part into sections which are connected end to end and have preset time length;
and summarizing the corresponding working condition information of each basic time period, and extracting the characteristics to obtain the use characteristics.
7. A mechanical component failure prediction apparatus, comprising:
the acquisition module is used for acquiring historical working condition information, and the historical working condition information comprises: operating condition information from when the mechanical component is started to be used until the current time;
the extraction module is used for extracting the use characteristics of the mechanical part based on the historical working condition information; the use characteristic is used for characterizing the use condition of the mechanical part;
the input module is used for obtaining a prediction result based on the use characteristic and the fault prediction model;
the fault prediction model is obtained by training a machine learning model which is set up in advance and is used for predicting whether the mechanical part fails within a preset time period from the current moment according to the use characteristics of the mechanical part.
8. The mechanical component failure prediction device of claim 7, further comprising: a training module for training the fault prediction model;
the training module is specifically configured to: obtaining a first preset number of sample use characteristics and a sample identification corresponding to each sample use characteristic as training samples; the sample identification is used for characterizing a mechanical part corresponding to the sample use characteristic, and whether a fault occurs in a target time period or not is judged; the target time period refers to a preset time period after the time corresponding to the sample historical working condition information of the sample use characteristics is extracted; and training the machine learning model based on the training samples to obtain the fault prediction model.
9. An electronic device, comprising:
a processor;
a memory for storing the processor-executable instructions;
the processor configured to perform the method of any of the preceding claims 1 to 6.
10. A computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, cause the processor to perform the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210479278.6A CN114881321A (en) | 2022-04-29 | 2022-04-29 | Mechanical component failure prediction method, device, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210479278.6A CN114881321A (en) | 2022-04-29 | 2022-04-29 | Mechanical component failure prediction method, device, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114881321A true CN114881321A (en) | 2022-08-09 |
Family
ID=82674360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210479278.6A Pending CN114881321A (en) | 2022-04-29 | 2022-04-29 | Mechanical component failure prediction method, device, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114881321A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117092933A (en) * | 2023-07-26 | 2023-11-21 | 天津通信广播集团有限公司 | Rotating machinery control method, apparatus, device and computer readable medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112149301A (en) * | 2020-09-23 | 2020-12-29 | 创新奇智(上海)科技有限公司 | Model training method, fault prediction method and device and electronic equipment |
| CN112346893A (en) * | 2020-11-10 | 2021-02-09 | 深圳市康必达控制技术有限公司 | Fault prediction method, device, terminal and storage medium |
| CN112529104A (en) * | 2020-12-23 | 2021-03-19 | 东软睿驰汽车技术（沈阳）有限公司 | Vehicle fault prediction model generation method, fault prediction method and device |
| WO2021109578A1 (en) * | 2019-12-02 | 2021-06-10 | 北京天元创新科技有限公司 | Method and apparatus for alarm prediction during service operation and maintenance, and electronic device |
-
2022
- 2022-04-29 CN CN202210479278.6A patent/CN114881321A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021109578A1 (en) * | 2019-12-02 | 2021-06-10 | 北京天元创新科技有限公司 | Method and apparatus for alarm prediction during service operation and maintenance, and electronic device |
| CN112149301A (en) * | 2020-09-23 | 2020-12-29 | 创新奇智(上海)科技有限公司 | Model training method, fault prediction method and device and electronic equipment |
| CN112346893A (en) * | 2020-11-10 | 2021-02-09 | 深圳市康必达控制技术有限公司 | Fault prediction method, device, terminal and storage medium |
| CN112529104A (en) * | 2020-12-23 | 2021-03-19 | 东软睿驰汽车技术（沈阳）有限公司 | Vehicle fault prediction model generation method, fault prediction method and device |
Non-Patent Citations (1)
| Title |
|---|
| 侯涛;: "机器学习技术在织机运行状况预测中的应用", 纺织科技进展, no. 09, 25 September 2018 (2018-09-25) * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117092933A (en) * | 2023-07-26 | 2023-11-21 | 天津通信广播集团有限公司 | Rotating machinery control method, apparatus, device and computer readable medium |
| CN117092933B (en) * | 2023-07-26 | 2024-04-16 | 天津通信广播集团有限公司 | Rotating machinery control method, apparatus, device and computer readable medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108897954A (en) | Wind turbines temperature pre-warning method and its system based on BootStrap confidence calculations | |
| CN110987439B (en) | Aeroengine fault prediction method based on Logitics regression and Xgboost model | |
| CN110688617B (en) | Fan vibration abnormity detection method and device | |
| CN112906969A (en) | Engine fault prediction method, engine fault prediction device, electronic equipment and storage medium | |
| CN110388315B (en) | Oil transfer pump fault identification method, device and system based on multi-source information fusion | |
| CN111538311B (en) | Flexible multi-state self-adaptive early warning method and device for mechanical equipment based on data mining | |
| CN112534370A (en) | System and method for predicting industrial machine failure | |
| CN110469461B (en) | Fracture estimation method and device for fan toothed belt and readable storage medium | |
| CN115239135A (en) | Construction risk identification management and control system and method | |
| CN114239734B (en) | Distributed vehicle-mounted health management system | |
| CN109507992B (en) | Method, device and equipment for predicting faults of locomotive brake system components | |
| CN115729761B (en) | Hard disk fault prediction method, system, equipment and medium | |
| CN107490479A (en) | Bearing residual life Forecasting Methodology and device | |
| CN114881321A (en) | Mechanical component failure prediction method, device, electronic device and storage medium | |
| CN115859791A (en) | Method and device for estimating residual life of aero-engine | |
| CN115650006A (en) | Elevator safety monitoring and early warning method and system based on big data | |
| CN114548280A (en) | Fault diagnosis model training method, fault diagnosis method and electronic equipment | |
| CN117312972A (en) | Method for identifying health state of scraper conveyor speed reducer | |
| CN116258482B (en) | Method for automatically selecting maintenance scheme, server and electronic equipment | |
| CN112524077A (en) | Method, device and system for detecting fan fault | |
| CN116910680A (en) | Remote fault detection method and system for fitness equipment | |
| CN110348005B (en) | Distribution network equipment state data processing method and device, computer equipment and medium | |
| CN114837902B (en) | Health degree evaluation method, system, equipment and medium for wind turbine generator | |
| CN113955149B (en) | Health diagnosis method and device for motor system | |
| CN115906437A (en) | Fan state determination method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |