CN113177650A - Predictive maintenance method and device for wagon compartment - Google Patents
Predictive maintenance method and device for wagon compartment Download PDFInfo
- Publication number
- CN113177650A CN113177650A CN202110594577.XA CN202110594577A CN113177650A CN 113177650 A CN113177650 A CN 113177650A CN 202110594577 A CN202110594577 A CN 202110594577A CN 113177650 A CN113177650 A CN 113177650A
- Authority
- CN
- China
- Prior art keywords
- maintenance
- fault
- health
- model
- data
- 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
- 238000012423 maintenance Methods 0.000 title claims abstract description 374
- 238000000034 method Methods 0.000 title claims abstract description 77
- 230000036541 health Effects 0.000 claims description 236
- 239000013598 vector Substances 0.000 claims description 136
- 238000003745 diagnosis Methods 0.000 claims description 63
- 238000012544 monitoring process Methods 0.000 claims description 56
- 238000012549 training Methods 0.000 claims description 37
- 238000005457 optimization Methods 0.000 claims description 35
- 238000007781 pre-processing Methods 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 23
- 238000012216 screening Methods 0.000 claims description 20
- 238000011161 development Methods 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 18
- 238000004458 analytical method Methods 0.000 claims description 11
- 230000004927 fusion Effects 0.000 claims description 10
- 230000006378 damage Effects 0.000 claims description 8
- 230000003862 health status Effects 0.000 claims description 8
- 230000002068 genetic effect Effects 0.000 claims description 7
- 239000002245 particle Substances 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 6
- 230000007423 decrease Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 230000007704 transition Effects 0.000 claims description 3
- 208000027418 Wounds and injury Diseases 0.000 claims description 2
- 238000012790 confirmation Methods 0.000 claims description 2
- 208000014674 injury Diseases 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 16
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 230000002159 abnormal effect Effects 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000013210 evaluation model Methods 0.000 description 3
- 230000003449 preventive effect Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 239000002699 waste 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
- G06Q10/20—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
- 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
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Train Traffic Observation, Control, And Security (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a predictive maintenance method and device for a railway wagon carriage. The method comprises the following steps: obtaining performance data of a railway freight car carriage; performing intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment; the rail wagon box is maintained according to a predictive maintenance scheme. According to the method and the device, intelligent decision is made according to the performance data, the maintenance decision influence factors and the maintenance strategy model, so that the optimal maintenance strategy of the wagon carriage is automatically obtained with high efficiency and high accuracy, maintenance is carried out based on the obtained predictive maintenance scheme, and low-cost operation and maintenance of the wagon carriage are achieved.
Description
Technical Field
The application relates to the technical field of rail transit, in particular to a method and a device for predictive maintenance of a railway wagon carriage.
Background
At present, the number of in-service rail transit transportation equipment in the whole country and the whole world is huge, and operation and maintenance service markets such as operation monitoring demands, maintenance demands and consumable updating demands with huge stocks are brought. With the continuous development of rail transit informatization, internet of things, industrial internet and big data analysis technology, intelligence becomes an important new direction for the development of rail transit transportation equipment in the world and becomes a new hotspot for the development of rail transit transportation equipment technology in China. In this case, in order to ensure the safety of the train, the investment in design and production, maintenance, monitoring and operation, and the like of the train needs to be increased. The problem that the accuracy of manual analysis of monitoring data and the prediction of manual faults in operation and maintenance is solved through a new technology becomes to be solved urgently at present.
At present, the aspect of making maintenance strategy decision based on real-time state analysis is blank at home and abroad, and the problems of low efficiency and high cost existing in manual (judging the state of a truck by a knocking means) or semi-manual (using a visual detection device to assist state inspection) maintenance strategy making in the transportation of the carriage of the existing railway truck still cannot be solved in the prior art. In addition, the accuracy and consistency of detection conclusions of the existing detection methods are difficult to guarantee, and the development of operation and maintenance intellectualization is greatly limited.
Disclosure of Invention
Based on the problems, the application provides a predictive maintenance method and device for a wagon compartment, so that a predictive maintenance decision for the wagon compartment can be automatically formed efficiently at low cost, the maintenance decision is more accurate, and the method and device are suitable for low-cost operation and maintenance of the wagon compartment.
The embodiment of the application discloses the following technical scheme:
in a first aspect of the present application, a method for predictive maintenance of a rail wagon box is provided, comprising:
obtaining performance data of a railway freight car carriage;
performing intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment;
maintaining the rail wagon box according to the predictive maintenance protocol;
the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities;
the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy;
the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
Optionally, the performance data comprises at least one of: health assessment information or fault prediction data.
Optionally, obtaining health assessment information for the rail wagon box comprises:
obtaining monitoring data of key components of the railway freight car;
processing the monitoring data to obtain a health index vector;
obtaining health assessment information for the key components of the rail wagon box using the health indicator vector and a health assessment model;
the health assessment model is obtained after training by utilizing a historical health index vector and a health state label of the historical health index vector; the historical health index vector is obtained by processing the historical data of the key component; the historical data is from the rail wagon car and/or from other rail wagon cars having the same key components as the rail wagon car.
Optionally, the processing the monitoring data to obtain a health indicator vector includes:
preprocessing the monitored real-time state signal, then extracting attenuation characteristic parameters, and screening and fusing characteristics to obtain the health index vector;
the historical health index vector is specifically a vector obtained by preprocessing the historical data of the key component, then extracting attenuation characteristic parameters, and screening and fusing the characteristics.
Optionally, the health status tag comprises: degree of damage and health grade of the component; the health assessment information includes: degree of injury and health grade.
Optionally, obtaining fault prediction data for the rail wagon box comprises:
acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on historical fault distribution statistics; the fault prediction model based on the historical fault distribution statistics is used for predicting the probability of faults of various types of faults of various time, mileage and lines in the future; or,
acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on time sequence state prediction; the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a period of time in the future.
Optionally, the fault prediction model based on time series state prediction includes: a fault diagnosis model and a performance trend prediction model;
the performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage;
and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
Optionally, the fault diagnosis model is established by:
obtaining historical fault data and historical fault mode information of each component of a railway wagon carriage;
obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model;
the signal tag includes: whether or not there is a failure, the location of the failure, and the type of failure.
Optionally, obtaining a health indicator vector according to the historical fault data includes:
and preprocessing historical fault data, extracting decline characteristic parameters, and performing characteristic screening and fusion to obtain a health index vector.
Optionally, the performance trend prediction model is built by:
the method comprises the steps of monitoring each part of a railway wagon carriage, and extracting an online state signal time sequence and an influence factor time sequence;
preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and training by using the health index vector time sequence and the influence factor time sequence to obtain the performance trend prediction model.
Optionally, the obtaining of the fault prediction data of the wagon box by using the monitoring data of each component of the wagon box and a fault prediction model based on time-series state prediction includes:
predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
obtaining a fault diagnosis result by using the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
judging whether the component has a fault according to the fault diagnosis result, and if so, outputting fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the preset time period is gradually increased, the health index after the preset time period is predicted through the performance trend prediction model, and a health index vector time sequence after the preset time period and subsequent steps are obtained.
Optionally, performing an intelligent decision according to the health assessment information, the fault prediction data, the maintenance decision influencing factor, and the maintenance policy model, including:
establishing a predictive maintenance behavior attribute set according to the health assessment information, the fault prediction data and the maintenance decision influencing factors;
according to the maintenance strategy model and all feasible predictive maintenance schemes, establishing a predictive maintenance behavior decision target set;
and evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the railway freight car according to a maximum fuzzy utility value principle.
Optionally, performing an intelligent decision according to the health assessment information, the fault prediction data, the maintenance decision influencing factor, and the maintenance policy model, including:
establishing a predictive maintenance strategy optimization model taking cost and availability as optimization targets;
the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved;
and solving by using the predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a pattern search method to obtain an optimized maintenance scheme serving as the predictive maintenance scheme of the wagon compartment.
Optionally, the method further includes:
and carrying out health state early warning according to the health assessment information and confirmation information provided by technicians.
Optionally, the method further includes:
and carrying out fault early warning according to the fault prediction data.
In a second aspect of the present application, there is provided a predictive maintenance device for a railway freight car, comprising:
the general acquisition module is used for acquiring performance data of the wagon compartment;
the intelligent decision module is used for carrying out intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment;
a maintenance module to perform maintenance on the rail wagon box according to the predictive maintenance protocol;
the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities;
the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy;
the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
Optionally, the total acquisition module comprises at least one of:
the system comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring health evaluation information of a railway wagon compartment;
and the second acquisition module is used for acquiring the fault prediction data of the wagon compartment.
Optionally, the first obtaining module includes:
the first monitoring data acquisition unit is used for acquiring monitoring data of key components of the railway wagon carriage;
the health index vector acquisition unit is used for processing the monitoring data to obtain a health index vector;
a health assessment unit for obtaining health assessment information of the key components of the rail wagon box using the health indicator vector and a health assessment model;
the health assessment model is obtained after training by utilizing a historical health index vector and a health state label of the historical health index vector; the historical health index vector is obtained by processing the historical data of the key component; the historical data is from the rail wagon car and/or from other rail wagon cars having the same key components as the rail wagon car.
Optionally, the health indicator vector obtaining unit is specifically configured to: preprocessing the monitored real-time state signal, then extracting attenuation characteristic parameters, and screening and fusing characteristics to obtain the health index vector; the historical health index vector is specifically a vector obtained by preprocessing the historical data of the key component, then extracting attenuation characteristic parameters, and screening and fusing the characteristics.
Optionally, the second obtaining module includes:
the first fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on historical fault distribution statistics; the fault prediction model based on the historical fault distribution statistics is used for predicting the probability of faults of various types of faults of various time, mileage and lines in the future; or,
the second fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on time sequence state prediction; the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a period of time in the future.
Optionally, the fault prediction model based on time series state prediction includes: a fault diagnosis model and a performance trend prediction model;
the performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage;
and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
Optionally, the fault diagnosis model is built by a model first training module, and the model first training module includes:
the system comprises a historical data acquisition unit, a fault analysis unit and a fault analysis unit, wherein the historical data acquisition unit is used for acquiring historical fault data and historical fault mode information of each component of a railway wagon carriage;
the first processing unit is used for obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
the first training unit is used for training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model;
the signal tag includes: whether or not there is a failure, the location of the failure, and the type of failure.
Optionally, the first processing unit is specifically configured to perform preprocessing on historical fault data, then extract regression characteristic parameters, perform characteristic screening and fusion, and obtain a health index vector.
Optionally, the performance trend prediction model is built by a model second training module, and the model second training module includes:
the monitoring data acquisition unit is used for extracting an online state signal time sequence and an influence factor time sequence by monitoring each part of a railway wagon carriage;
the second processing unit is used for preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and the second training unit is used for training to obtain the performance trend prediction model by utilizing the health index vector time sequence and the influence factor time sequence.
Optionally, the failed second prediction unit comprises:
the health index vector time sequence prediction unit is used for predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
the fault diagnosis unit is used for obtaining a fault diagnosis result by utilizing the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
the fault judging unit is used for judging whether the component has a fault according to the fault diagnosis result, and if so, the data output unit outputs fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the counting unit carries out incremental transition on the preset time period, and the health index vector time series prediction unit executes corresponding functions again.
Optionally, the intelligent decision module includes:
a first establishing unit, configured to establish a predictive maintenance behavior attribute set according to the health assessment information, the fault prediction data, and the maintenance decision influencing factors;
the second establishing unit is used for establishing a predictive maintenance behavior decision target set according to the maintenance strategy model and summarizing feasible predictive maintenance schemes;
and the reasoning decision unit is used for evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the wagon compartment according to a maximum fuzzy utility value principle.
Optionally, the intelligent decision module includes:
a third establishing unit, for establishing a predictive maintenance strategy optimization model with cost and availability as optimization targets;
the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved;
and the optimization decision unit is used for solving by utilizing the predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a mode search method to obtain an optimized maintenance scheme serving as the predictive maintenance scheme of the wagon compartment.
Compared with the prior art, the method has the following beneficial effects:
the application provides a predictive maintenance method for a railway freight car, which comprises the following steps: obtaining performance data of a railway freight car carriage; performing intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment; maintaining the railway freight car according to a predictive maintenance scheme; the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities; the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy; the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan. According to the method and the device, intelligent decision is made according to the performance data, the maintenance decision influence factors and the maintenance strategy model, so that the optimal maintenance strategy of the wagon carriage is automatically obtained with high efficiency and high accuracy, maintenance is carried out based on the obtained predictive maintenance scheme, and low-cost operation and maintenance of the wagon carriage are achieved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a flow chart of a method for predictive maintenance of a rail wagon box according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram illustrating a process of establishing and applying a health assessment model according to an embodiment of the present disclosure;
fig. 3 is a schematic flow chart illustrating the establishment of a fault prediction related model and a fault prediction process according to an embodiment of the present application;
FIG. 4 is a block diagram of an implementation of a method for predictive maintenance of a rail wagon box according to an embodiment of the present disclosure;
FIG. 5 is a schematic diagram illustrating an application of a multi-objective optimization model according to an embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of a predictive maintenance device for a railway wagon box according to an embodiment of the present application.
Detailed Description
As previously described, current rail wagon car maintenance strategies typically require manual or semi-manual development, which is inefficient and costly. In addition, the accuracy and consistency of the detection conclusions are difficult to guarantee, which results in that maintenance strategies may not be applicable. In view of the above problems, the inventors have studied to provide a method and apparatus for predictive maintenance of a railway freight car. In the method, the intelligent decision of the railway freight car predictive maintenance scheme is realized by combining the performance data of the railway freight car, the maintenance decision influence factors and the maintenance strategy model, and the optimal predictive maintenance scheme can be obtained finally. The method automatically and efficiently obtains the cost of the predictive maintenance scheme, does not need manual or semi-manual participation, and has higher accuracy and applicability.
In order to make the technical solutions of the present application better understood, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, 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 application.
The method comprises the following steps:
referring to fig. 1, a flowchart of a method for predictive maintenance of a railway freight car according to an embodiment of the present application is shown. The method of predictive maintenance of a railway freight car as shown in fig. 1 comprises:
step 101: performance data of the rail wagon box is obtained.
The performance data herein can reflect the current or future performance of the components in the railcar car. As an alternative implementation, the performance data includes at least one of health assessment information and fault prediction data.
An example implementation of obtaining health assessment information for a rail wagon box is first described. In this example implementation, the obtaining of the cabin health assessment information is accomplished by a health assessment model.
The health assessment model is introduced first. In the embodiment of the application, a health assessment model is obtained by training with a historical health indicator vector and a health status label of the historical health indicator vector. The historical health index vector is obtained by processing the historical data of the key components; the historical data is from rail wagons and/or from other rail wagons having the same key components as the rail wagons.
Fig. 2 is a schematic diagram of the establishment process and the application process of the health assessment model. As shown in fig. 2, the historical data of different health states accumulated in the railway wagon compartment are subjected to signal preprocessing, and decline characteristic parameters are extracted. And obtaining a representative health index after characteristic screening and fusion. As an example, for a sensor signal with waveform characteristics (such as a car vibration signal), characteristic parameters that better reflect the time-series variation trend of performance, such as Root Mean Square (RMS), wavelet packet energy, Hilbert envelope spectrum bandwidth energy, spectral kurtosis, and the like, may be selected. In addition, other time domain, frequency domain and time-frequency domain parameters can be integrated to establish multi-dimensional characteristic parameters, and then redundant characteristics and irrelevant characteristics are eliminated and the characteristic dimension is reduced by utilizing characteristic screening methods such as an expert system method, a Filter method, a variance value division method, a Fisher value division method and the like; meanwhile, the main elements of the health state feature vector can be extracted and the feature vector can be fused by using feature fusion algorithms such as Principal Component Analysis (PCA), Transfer Component Analysis (TCA) and the like, so that the feature dimension is further reduced, and finally, the health index vector which is rich in state information and small in dimension is obtained. These health indicator vectors prepared for training the health assessment model are collectively referred to as historical health indicator vectors.
To train the health assessment model, a health status label of the historical health indicator vector is also prepared. The health status label to which the historical health indicator vector corresponds is essentially an identification of the health status of the device providing the historical data. In an example implementation manner, the historical overhaul report of the equipment can be combined, and based on expert experience and a physical model, the division and identification of the historical damage degree, the health level and the like of each component corresponding to historical data can be completed, so that a health state label can be obtained. For example, aiming at a key rotating mechanism of a railway wagon carriage, the process from an initial working state to complete function loss of equipment is divided into different degradation states through a discrete health state degradation process, and various performance parameter values can be comprehensively measured on the basis of performance parameter information such as the abrasion loss, the crack length and the like of the mechanism obtained through maintenance, so that a health state label is obtained. The health status label includes: degree of damage and health of the component.
After obtaining the health index vectors of the key components of each historical carriage and the corresponding health state labels thereof, the health evaluation model of each key component of the railway wagon carriage can be trained based on Machine learning/deep learning algorithms such as a Support Vector Machine (SVM), a Neural network, a Convolutional Neural Network (CNN), Latent Dirichlet Allocation (LDA) and the like, so as to realize the establishment of the health evaluation model. Finally, the well-established health assessment model may be saved in a database.
The process of applying the health assessment model is described below, with continued reference to FIG. 2. Firstly, obtaining monitoring data of key components of a railway wagon carriage, wherein the monitoring data can be real-time monitoring data, and processing the monitoring data to obtain a health index vector. Specifically, the monitored real-time state signal may be preprocessed, and then attenuation characteristic parameters are extracted and characteristic screening and fusion are performed to obtain the health index vector. And obtaining the health evaluation information of the key components of the railway wagon carriage according to the health index vector obtained according to the input monitoring data through the health evaluation model. The health assessment information corresponds to the aforementioned health status labels used to train the health assessment model. The health assessment information may include: degree of damage and health level of critical components.
One example implementation of obtaining fault prediction data for a rail wagon box is next described. In this example implementation, the obtaining of the car fault prediction data is accomplished by a fault prediction model.
A fault prediction model is introduced. In the embodiment of the present application, as an example, a fault prediction model based on historical fault distribution statistics may be adopted.
The fault prediction model based on historical fault distribution statistics is mainly used for predicting the probability of occurrence of various faults of various time, mileage and lines by analyzing the distribution conditions of various historical fault types along with time, mileage and lines, for example, the probability distribution of abnormal longitudinal force of a coupler, abnormal pressure of a brake cylinder, abnormal stroke of the brake cylinder, abnormal pressure of a brake pipe, abnormal vibration of a vehicle body and the like under different driving mileage, time and different lines is obtained through analysis, so that the time, mileage, maintenance objects, key maintenance lines and the like of predictive maintenance are reasonably arranged. Since the statistical analysis method is very mature and commonly used, it is not repeated here. And the fault prediction model based on historical fault distribution statistics is used for predicting the probability of faults of various types of future time, mileage and lines. When the fault prediction model based on the historical fault distribution statistics is applied, the fault prediction data of the railway wagon compartment can be obtained by utilizing the monitoring data of all parts of the railway wagon compartment and the fault prediction model based on the historical fault distribution statistics.
As another example, a fault prediction model based on timing state prediction may also be employed.
The fault prediction model based on time sequence state prediction mainly predicts the development trend of the future state of the vehicle based on a performance trend prediction model (also called as a time sequence prediction model) according to the historical running state data of the vehicle, and identifies the fault state and the fault type corresponding to the future state parameter by combining a fault diagnosis model, thereby predicting the occurrence of the future fault. Therefore, in the embodiment of the application, the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a future period of time. When the fault prediction model based on the time sequence state prediction is applied, the fault prediction data of the wagon compartment can be obtained by utilizing the monitoring data of all parts of the wagon compartment and the fault prediction model based on the time sequence state prediction.
Fig. 3 is a schematic flow chart illustrating the building of a fault prediction related model and a fault prediction process according to an embodiment of the present application. The fault prediction model comprises two parts, namely a fault diagnosis model and a performance trend prediction model. The performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage; and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
As shown in fig. 3, the fault diagnosis model needs to obtain a health indicator vector and a signal label. The health index vector is obtained by preprocessing data and signals in a rail transit transportation equipment database, extracting attenuation characteristic parameters, and performing characteristic screening and fusion. The signal tag is also marked according to data and signals in a rail transit transportation equipment database, and the signal tag can comprise: failure or not, failure location, and failure type. And training a fault diagnosis model based on the health index vector and the signal label. The fault diagnosis model may be an SVM, CNN, or LDA model.
The fault diagnosis model is established by the following operations:
obtaining historical fault data and historical fault mode information of each component of a railway wagon carriage;
obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
and training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model.
Obtaining a health index vector according to the historical fault data, wherein the obtaining of the health index vector comprises the following steps:
and preprocessing historical fault data, extracting decline characteristic parameters, and performing characteristic screening and fusion to obtain a health index vector.
For the establishment of a performance trend prediction model, an online state signal time sequence and an influence factor time sequence need to be extracted from online monitoring data of all parts of rail transit transportation equipment. And preprocessing the online state signal time sequence, then obtaining a health index vector, and extracting to obtain a health index vector time sequence. And training to obtain a performance trend prediction model by using the health index vector time sequence and the influence factor time sequence. When fault prediction is carried out, the performance trend prediction model predicts the health index after the time period n delta t to obtain a health index vector time sequence in the future n delta t time. The time sequence is provided to a fault diagnosis model, a fault anchor result is given by the fault diagnosis model, the fault anchor result comprises a fault label of each part in a future period, and whether the part has a fault or not can be judged based on the label. If the fault is detected, the future fault time point, the future fault position and the future fault type can be obtained, so that fault early warning can be implemented, and the decision of a predictive maintenance scheme can be made. If there is no fault, the original n can be added by 1, so that the value of n is increased.
The performance trend prediction model is established by the following operations:
the method comprises the steps of monitoring each part of a railway wagon carriage, and extracting an online state signal time sequence and an influence factor time sequence;
preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and training by using the health index vector time sequence and the influence factor time sequence to obtain the performance trend prediction model.
The method for obtaining the fault prediction data of the wagon compartment by using the monitoring data of each component of the wagon compartment and the fault prediction model based on time sequence state prediction comprises the following steps:
predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
obtaining a fault diagnosis result by using the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
judging whether the component has a fault according to the fault diagnosis result, and if so, outputting fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the preset time period is gradually increased, the health index after the preset time period is predicted through the performance trend prediction model, and a health index vector time sequence after the preset time period and subsequent steps are obtained.
Through the step 101, the acquisition of the health assessment information and the fault prediction data of the wagon compartment is completed.
Step 102: and carrying out intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment.
In an embodiment of the present application, the maintenance decision influencing factors include at least one of: security, time, cost, availability, job plan, maintenance resources, or maintenance capabilities. These considerations of maintenance decision influencing factors all influence the decision of the predictive maintenance solution.
In an embodiment of the present application, the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy. The strategies in the maintenance strategy model are the basis for forming a predictive maintenance solution.
Service life based maintenance strategy: according to the fault prediction result based on statistical analysis, when the accumulated running time and mileage of the rail transit transportation equipment reach a certain value, the parts with higher fault distribution frequency corresponding to the time and the mileage are scheduled to be overhauled and maintained in advance.
Opportunity maintenance strategy: when a certain part of the rail transit transportation equipment is subjected to maintenance decision analysis, the influence effect between the part and the related part is comprehensively considered. The maintenance work of one part of the vehicle also means a maintenance opportunity of the relevant part having a dependency relationship therewith. The opportunity maintenance strategy can be configured according to different structures of the system, various parts are utilized to maintain the shutdown opportunities, and maintenance schemes of all rail transit transportation equipment of operators are dynamically scheduled and optimized through a cost balance theory.
And (3) a visual maintenance strategy: scheduling preventive maintenance for the potential fault in advance of the predicted fault time point according to fault prediction data and health state information of a fault prediction model based on time sequence state prediction; meanwhile, temporary maintenance is arranged according to situations by considering a vehicle transportation operation plan, maintenance personnel time arrangement and the like.
Group maintenance strategy: the parts of the same type running under the same condition in the rail transit transportation equipment can be divided into a group, and if one of the parts fails or is maintained, maintenance operation is carried out on the parts of the same type in the same group which do not fail.
Preventive maintenance periodic maintenance adjustment strategy: adjusting an original regular maintenance plan according to real-time health state information and fault prediction data of each vehicle of the rail transit transportation equipment, for example, shortening a maintenance period along with the deterioration of the health state; and according to the predicted future fault category, newly adding corresponding key maintenance items in a future maintenance plan and the like.
It can be said that when the performance data includes health assessment information and fault prediction data, the health assessment information, fault prediction data, maintenance decision influencing factors and maintenance strategy models guide the resulting predictive maintenance schedule for the rail wagon box. The predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
Two alternative implementations of step 102 are provided below.
In an alternative implementation, step 102 is implemented based on an inference decision method when executed. Specifically, a predictive maintenance behavior attribute set can be established according to the health assessment information, the fault prediction data and the maintenance decision influencing factors; establishing a predictive maintenance behavior decision target set according to a maintenance strategy model and summarizing feasible predictive maintenance schemes; and evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the wagon compartment according to the maximum fuzzy utility value principle.
In another alternative implementation, step 102 is implemented based on an optimization decision method when executed. Specifically, a predictive maintenance strategy optimization model with cost and availability as optimization objectives can be established; the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved; and solving by using a predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a mode search method to obtain an optimized maintenance scheme serving as a predictive maintenance scheme of the wagon compartment. Therefore, the above parameters to be solved are included in the predictive maintenance scheme.
Of course, the above inference decision-based and optimization decision-based schemes can be implemented in parallel, and finally a predictive maintenance scheme is selected from the results of the two decisions. The above schemes based on reasoning decision and optimization decision are examples, and in practical application, intelligent decisions in other ways can be made based on health assessment information, fault prediction data, maintenance decision influencing factors and a maintenance strategy model, so that a predictive maintenance scheme for a railway freight car is finally obtained. Therefore, the specific implementation of step 102 is not limited herein.
Step 103: the rail wagon box is maintained according to a predictive maintenance scheme.
In this step, the railway freight car may be maintained according to the predictive maintenance schedule obtained in step 102.
In the method, when the performance data comprises health evaluation information and fault prediction data, intelligent decision is made according to the health evaluation information, the fault prediction data, the maintenance decision influence factors and the maintenance strategy model, so that the optimal maintenance strategy of the wagon carriage is automatically obtained with high efficiency and high accuracy, maintenance is carried out based on the obtained predictive maintenance scheme, and low-cost operation and maintenance of the wagon carriage are realized.
In the embodiment of the present application, as shown in fig. 2, after obtaining the health assessment information, a technician may manually confirm the assessment information, and then perform a health state early warning according to the health assessment information when it is confirmed that the damage degree indicated by the assessment information is severe or the health level of the key component is poor. In practical application, the early warning modes include multiple modes, for example, alarm of ringing or early warning by sending information to operation and maintenance personnel, and the like, so the early warning modes are not limited. In addition, the method can also carry out fault early warning according to the fault prediction data.
According to the method and the device, the railway wagon carriage fault can be accurately predicted before the occurrence of the railway wagon carriage fault, and effective measures can be taken to prevent the occurrence of the fault. Through preventive maintenance, the maintenance period of the carriage component is within a controllable range, the schedule arrangement of an equipment maintenance team is optimized, and the working efficiency is improved. In addition, according to the technical scheme provided by the embodiment of the application, a predictive maintenance scheme is obtained, and waste of spare parts can be avoided. For example, the purchase amount of spare parts can be determined according to a predictive maintenance scheme, so that the production cost is further saved; through a predictive maintenance scheme, the added value of products (such as key parts of a carriage) is further improved, the outage time of the wagon carriage can be greatly reduced, and the cost loss caused by the outage is reduced.
Fig. 4 is a block diagram of an implementation of a method for predictive maintenance of a rail wagon according to an embodiment of the present application. It should be noted that the monitoring data used in the use of the health assessment model and the fault prediction model may be direct monitoring data or indirect monitoring data.
In addition, the application also provides a hierarchical predictive maintenance decision method for the wagon compartment based on the health state division. Referring to fig. 5, a schematic diagram of an application of the multi-objective optimization model is shown. The model can realize the hierarchical predictive maintenance decision of the railway freight car. And dividing the health state based on the future fault mode of the health state of each component of the carriage and the correlation analysis result of each component and the overall fault. And then performing hierarchical predictive maintenance according to the results of the optimization and the health state division of each level of maintenance strategy. The maintenance strategies of each level can be realized by genetic algorithm or particle swarm optimization and the like. The decision content of the hierarchical predictive maintenance comprises: maintenance mode, maintenance target, maintenance timing, maintenance cycle, and the like. In the embodiment of the application, the multi-objective optimization model is optimized based on the availability, the cost, the risk and the like as targets.
Based on the predictive maintenance method for the wagon box provided by the foregoing embodiment, correspondingly, the present application also provides a predictive maintenance device for the wagon box. The implementation of the device is described below with reference to the embodiments and the drawings.
Device embodiment
Referring to fig. 6, a schematic diagram of a predictive maintenance device for a railway freight car is provided. As shown in fig. 6, the predictive maintenance device 600 for a railway freight car includes:
a general acquisition module 601, configured to acquire performance data of a wagon compartment;
an intelligent decision module 602, configured to perform an intelligent decision according to the performance data, the maintenance decision influencing factor, and the maintenance policy model, so as to obtain a predictive maintenance scheme for the wagon box;
a maintenance module 603 for performing maintenance on the rail wagon box according to the predictive maintenance schedule;
the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities;
the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy;
the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
According to the method and the device, intelligent decision is made according to the performance data, the maintenance decision influence factors and the maintenance strategy model, so that the optimal maintenance strategy of the wagon carriage is automatically obtained with high efficiency and high accuracy, maintenance is carried out based on the obtained predictive maintenance scheme, and low-cost operation and maintenance of the wagon carriage are achieved.
In a possible implementation manner, the total obtaining module 601 includes at least one of the following:
the system comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring health evaluation information of a railway wagon compartment;
and the second acquisition module is used for acquiring the fault prediction data of the wagon compartment.
In a possible implementation manner, the first obtaining module includes:
the first monitoring data acquisition unit is used for acquiring monitoring data of key components of the railway wagon carriage;
the health index vector acquisition unit is used for processing the monitoring data to obtain a health index vector;
a health assessment unit for obtaining health assessment information of the key components of the rail wagon box using the health indicator vector and a health assessment model;
the health assessment model is obtained after training by utilizing a historical health index vector and a health state label of the historical health index vector; the historical health index vector is obtained by processing the historical data of the key component; the historical data is from the rail wagon car and/or from other rail wagon cars having the same key components as the rail wagon car.
In a possible implementation manner, the health indicator vector obtaining unit is specifically configured to: preprocessing the monitored real-time state signal, then extracting attenuation characteristic parameters, and screening and fusing characteristics to obtain the health index vector; the historical health index vector is specifically a vector obtained by preprocessing the historical data of the key component, then extracting attenuation characteristic parameters, and screening and fusing the characteristics.
In a possible implementation manner, the second obtaining module includes:
the first fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on historical fault distribution statistics; the fault prediction model based on the historical fault distribution statistics is used for predicting the probability of faults of various types of faults of various time, mileage and lines in the future; or,
the second fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on time sequence state prediction; the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a period of time in the future.
In one possible implementation, the fault prediction model based on time series state prediction includes: a fault diagnosis model and a performance trend prediction model;
the performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage;
and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
In one possible implementation, the fault diagnosis model is built by a model first training module, and the model first training module includes:
the system comprises a historical data acquisition unit, a fault analysis unit and a fault analysis unit, wherein the historical data acquisition unit is used for acquiring historical fault data and historical fault mode information of each component of a railway wagon carriage;
the first processing unit is used for obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
the first training unit is used for training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model;
the signal tag includes: whether or not there is a failure, the location of the failure, and the type of failure.
In a possible implementation manner, the first processing unit is specifically configured to perform preprocessing on historical fault data, then extract regression characteristic parameters, and perform characteristic screening and fusion to obtain a health index vector.
In one possible implementation, the performance trend prediction model is built by a model second training module, and the model second training module includes:
the monitoring data acquisition unit is used for extracting an online state signal time sequence and an influence factor time sequence by monitoring each part of a railway wagon carriage;
the second processing unit is used for preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and the second training unit is used for training to obtain the performance trend prediction model by utilizing the health index vector time sequence and the influence factor time sequence.
In one possible implementation, the failed second prediction unit includes:
the health index vector time sequence prediction unit is used for predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
the fault diagnosis unit is used for obtaining a fault diagnosis result by utilizing the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
the fault judging unit is used for judging whether the component has a fault according to the fault diagnosis result, and if so, the data output unit outputs fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the counting unit carries out incremental transition on the preset time period, and the health index vector time series prediction unit executes corresponding functions again.
In one possible implementation manner, the intelligent decision module 602 includes:
a first establishing unit, configured to establish a predictive maintenance behavior attribute set according to the health assessment information, the fault prediction data, and the maintenance decision influencing factors;
the second establishing unit is used for establishing a predictive maintenance behavior decision target set according to the maintenance strategy model and summarizing feasible predictive maintenance schemes;
and the reasoning decision unit is used for evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the wagon compartment according to a maximum fuzzy utility value principle.
In one possible implementation manner, the intelligent decision module 602 includes:
a third establishing unit, for establishing a predictive maintenance strategy optimization model with cost and availability as optimization targets;
the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved;
and the optimization decision unit is used for solving by utilizing the predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a mode search method to obtain an optimized maintenance scheme serving as the predictive maintenance scheme of the wagon compartment.
It should be noted that, in the present specification, all the embodiments are described in a progressive manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus and system embodiments, since they are substantially similar to the method embodiments, they are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described embodiments of the apparatus and system are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts suggested as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
The above description is only one specific embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (27)
1. A method of predictive maintenance of a railway freight car, comprising:
obtaining performance data of a railway freight car carriage;
performing intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment;
maintaining the rail wagon box according to the predictive maintenance protocol;
the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities;
the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy;
the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
2. The method of claim 1, wherein the performance data comprises at least one of: health assessment information or fault prediction data.
3. The method of claim 2, wherein obtaining health assessment information for a rail wagon box comprises:
obtaining monitoring data of key components of the railway freight car;
processing the monitoring data to obtain a health index vector;
obtaining health assessment information for the key components of the rail wagon box using the health indicator vector and a health assessment model;
the health assessment model is obtained after training by utilizing a historical health index vector and a health state label of the historical health index vector; the historical health index vector is obtained by processing the historical data of the key component; the historical data is from the rail wagon car and/or from other rail wagon cars having the same key components as the rail wagon car.
4. The method of claim 3, wherein the processing the monitoring data to obtain a health indicator vector comprises:
preprocessing the monitored real-time state signal, then extracting attenuation characteristic parameters, and screening and fusing characteristics to obtain the health index vector;
the historical health index vector is specifically a vector obtained by preprocessing the historical data of the key component, then extracting attenuation characteristic parameters, and screening and fusing the characteristics.
5. The method of claim 3, wherein the health status tag comprises: degree of damage and health grade of the component; the health assessment information includes: degree of injury and health grade.
6. The method of claim 2, wherein obtaining fault prediction data for a rail wagon box comprises:
acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on historical fault distribution statistics; the fault prediction model based on the historical fault distribution statistics is used for predicting the probability of faults of various types of faults of various time, mileage and lines in the future; or,
acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on time sequence state prediction; the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a period of time in the future.
7. The method of claim 6, wherein the time series state prediction based fault prediction model comprises: a fault diagnosis model and a performance trend prediction model;
the performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage;
and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
8. The method of claim 7, wherein the fault diagnosis model is established by:
obtaining historical fault data and historical fault mode information of each component of a railway wagon carriage;
obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model;
the signal tag includes: whether or not there is a failure, the location of the failure, and the type of failure.
9. The method of claim 8, wherein deriving a health indicator vector from the historical fault data comprises:
and preprocessing historical fault data, extracting decline characteristic parameters, and performing characteristic screening and fusion to obtain a health index vector.
10. The method of claim 8, wherein the performance trend prediction model is created by:
the method comprises the steps of monitoring each part of a railway wagon carriage, and extracting an online state signal time sequence and an influence factor time sequence;
preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and training by using the health index vector time sequence and the influence factor time sequence to obtain the performance trend prediction model.
11. The method of claim 10, wherein said obtaining fault prediction data for said rail wagon box using monitored data for components of said rail wagon box and a fault prediction model based on time series state predictions comprises:
predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
obtaining a fault diagnosis result by using the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
judging whether the component has a fault according to the fault diagnosis result, and if so, outputting fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the preset time period is gradually increased, the health index after the preset time period is predicted through the performance trend prediction model, and a health index vector time sequence after the preset time period and subsequent steps are obtained.
12. The method of claim 2, wherein making an intelligent decision based on the health assessment information, the fault prediction data, maintenance decision influencing factors, and a maintenance policy model comprises:
establishing a predictive maintenance behavior attribute set according to the health assessment information, the fault prediction data and the maintenance decision influencing factors;
according to the maintenance strategy model and all feasible predictive maintenance schemes, establishing a predictive maintenance behavior decision target set;
and evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the railway freight car according to a maximum fuzzy utility value principle.
13. The method of claim 2, wherein making an intelligent decision based on the health assessment information, the fault prediction data, maintenance decision influencing factors, and a maintenance policy model comprises:
establishing a predictive maintenance strategy optimization model taking cost and availability as optimization targets;
the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved;
and solving by using the predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a pattern search method to obtain an optimized maintenance scheme serving as the predictive maintenance scheme of the wagon compartment.
14. The method of claim 2, further comprising:
and carrying out health state early warning according to the health assessment information and confirmation information provided by technicians.
15. The method of claim 6, further comprising:
and carrying out fault early warning according to the fault prediction data.
16. A predictive maintenance device for a railway freight car, comprising:
the general acquisition module is used for acquiring performance data of the wagon compartment;
the intelligent decision module is used for carrying out intelligent decision according to the performance data, the maintenance decision influence factors and the maintenance strategy model to obtain a predictive maintenance scheme for the wagon compartment;
a maintenance module to perform maintenance on the rail wagon box according to the predictive maintenance protocol;
the maintenance decision influencing factors include at least one of: security, time, cost, availability, job planning, maintenance resources or maintenance capabilities;
the maintenance policy model includes at least one of the following policies: a service-age based maintenance policy, an opportunity maintenance policy, a look-at-sight maintenance policy, a group maintenance policy, or a preventative maintenance cycle adjustment policy;
the predictive maintenance protocol includes at least one of the following components: maintenance mode, maintenance type, maintenance content, maintenance opportunity or spare part order plan.
17. The apparatus of claim 16, wherein the overall acquisition module comprises at least one of:
the system comprises a first acquisition module, a second acquisition module and a control module, wherein the first acquisition module is used for acquiring health evaluation information of a railway wagon compartment;
and the second acquisition module is used for acquiring the fault prediction data of the wagon compartment.
18. The apparatus of claim 17, wherein the first obtaining module comprises:
the first monitoring data acquisition unit is used for acquiring monitoring data of key components of the railway wagon carriage;
the health index vector acquisition unit is used for processing the monitoring data to obtain a health index vector;
a health assessment unit for obtaining health assessment information of the key components of the rail wagon box using the health indicator vector and a health assessment model;
the health assessment model is obtained after training by utilizing a historical health index vector and a health state label of the historical health index vector; the historical health index vector is obtained by processing the historical data of the key component; the historical data is from the rail wagon car and/or from other rail wagon cars having the same key components as the rail wagon car.
19. The apparatus according to claim 18, wherein the health indicator vector obtaining unit is specifically configured to: preprocessing the monitored real-time state signal, then extracting attenuation characteristic parameters, and screening and fusing characteristics to obtain the health index vector; the historical health index vector is specifically a vector obtained by preprocessing the historical data of the key component, then extracting attenuation characteristic parameters, and screening and fusing the characteristics.
20. The apparatus of claim 17, wherein the second obtaining module comprises:
the first fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on historical fault distribution statistics; the fault prediction model based on the historical fault distribution statistics is used for predicting the probability of faults of various types of faults of various time, mileage and lines in the future; or,
the second fault prediction unit is used for acquiring fault prediction data of the wagon compartment by utilizing monitoring data of all parts of the wagon compartment and a fault prediction model based on time sequence state prediction; the fault prediction model based on the time sequence state prediction is used for predicting the development trend of each type of fault in a period of time in the future.
21. The apparatus of claim 20, wherein the time series state prediction based fault prediction model comprises: a fault diagnosis model and a performance trend prediction model;
the performance trend prediction model is used for predicting the development trend of the future state of the railway wagon carriage according to the monitoring data of all the components of the railway wagon carriage;
and the fault diagnosis model is used for identifying the fault state and the fault type corresponding to the future state parameters so as to predict the occurrence of the future fault.
22. The apparatus of claim 21, wherein the fault diagnosis model is built by a model first training module comprising:
the system comprises a historical data acquisition unit, a fault analysis unit and a fault analysis unit, wherein the historical data acquisition unit is used for acquiring historical fault data and historical fault mode information of each component of a railway wagon carriage;
the first processing unit is used for obtaining a health index vector and a signal label corresponding to the health index vector according to the historical fault data and the historical fault mode information;
the first training unit is used for training by using the health index vector and the signal label corresponding to the health index vector to obtain the fault diagnosis model;
the signal tag includes: whether or not there is a failure, the location of the failure, and the type of failure.
23. The apparatus according to claim 22, wherein the first processing unit is specifically configured to preprocess historical fault data, extract regression feature parameters, and perform feature screening and fusion to obtain the health indicator vector.
24. The apparatus of claim 22, wherein the performance trend prediction model is built by a model second training module comprising:
the monitoring data acquisition unit is used for extracting an online state signal time sequence and an influence factor time sequence by monitoring each part of a railway wagon carriage;
the second processing unit is used for preprocessing the online state signal time sequence and extracting a health index vector to obtain a health index vector time sequence;
and the second training unit is used for training to obtain the performance trend prediction model by utilizing the health index vector time sequence and the influence factor time sequence.
25. The apparatus of claim 24, wherein the faulty second prediction unit comprises:
the health index vector time sequence prediction unit is used for predicting the health index after a preset time period through the performance trend prediction model to obtain a health index vector time sequence after the preset time period;
the fault diagnosis unit is used for obtaining a fault diagnosis result by utilizing the fault diagnosis model and the fault diagnosis model; the fault diagnosis result comprises fault labels of all components in a future period of time;
the fault judging unit is used for judging whether the component has a fault according to the fault diagnosis result, and if so, the data output unit outputs fault prediction data; the fault prediction data includes: future fault time points, future fault locations, and future fault types; if not, the counting unit carries out incremental transition on the preset time period, and the health index vector time series prediction unit executes corresponding functions again.
26. The apparatus of claim 17, wherein the intelligent decision module comprises:
a first establishing unit, configured to establish a predictive maintenance behavior attribute set according to the health assessment information, the fault prediction data, and the maintenance decision influencing factors;
the second establishing unit is used for establishing a predictive maintenance behavior decision target set according to the maintenance strategy model and summarizing feasible predictive maintenance schemes;
and the reasoning decision unit is used for evaluating each predictive maintenance scheme by using a fuzzy reasoning or evidence reasoning method according to the predictive maintenance behavior attribute set and the predictive maintenance behavior decision target set, and determining the optimal predictive maintenance scheme for the wagon compartment according to a maximum fuzzy utility value principle.
27. The apparatus of claim 17, wherein the intelligent decision module comprises:
a third establishing unit, for establishing a predictive maintenance strategy optimization model with cost and availability as optimization targets;
the predictive maintenance strategy optimization model takes maintenance resources, maintenance capacity and an operation plan as constraint conditions, and takes a maintenance period, maintenance opportunity and a maintenance mode as parameters to be solved;
and the optimization decision unit is used for solving by utilizing the predictive maintenance strategy optimization model through a genetic algorithm, a particle swarm algorithm or a mode search method to obtain an optimized maintenance scheme serving as the predictive maintenance scheme of the wagon compartment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110594577.XA CN113177650A (en) | 2021-05-28 | 2021-05-28 | Predictive maintenance method and device for wagon compartment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110594577.XA CN113177650A (en) | 2021-05-28 | 2021-05-28 | Predictive maintenance method and device for wagon compartment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113177650A true CN113177650A (en) | 2021-07-27 |
Family
ID=76927136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110594577.XA Pending CN113177650A (en) | 2021-05-28 | 2021-05-28 | Predictive maintenance method and device for wagon compartment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113177650A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113792944A (en) * | 2021-11-16 | 2021-12-14 | 深圳普菲特信息科技股份有限公司 | Predictive maintenance method and system |
| CN113988326A (en) * | 2021-10-09 | 2022-01-28 | 南京理工大学 | Subway equipment maintenance optimization method and system |
| CN114707918A (en) * | 2021-12-07 | 2022-07-05 | 西南交通大学 | Railway hierarchical node transportation method |
| CN115034306A (en) * | 2022-06-13 | 2022-09-09 | 昆明理工大学 | Aluminum electrolysis fault prediction and safe operation and maintenance method based on extensible neural network |
| CN117152930A (en) * | 2023-08-04 | 2023-12-01 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway cargo transportation state early warning method and device and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050060251A (en) * | 2003-12-16 | 2005-06-22 | 한국철도기술연구원 | Prevention system and the method for maintenance of railway car using rams analysis |
| CN108520321A (en) * | 2018-04-03 | 2018-09-11 | 广西大学 | A kind of train bogie critical component preventative maintenance decision optimization model based on minimax ant group algorithm |
| CN109886485A (en) * | 2019-02-13 | 2019-06-14 | 上海大学 | A kind of traffic infrastructure maintenance policy optimization method |
| CN109934358A (en) * | 2019-01-30 | 2019-06-25 | 中国人民解放军32181部队 | Equipment failure prediction and health evaluating method, system and terminal device |
| CN109948169A (en) * | 2017-12-20 | 2019-06-28 | 中国中车股份有限公司 | A kind of railway freight-car prognostic and health management system |
| CN110232452A (en) * | 2019-06-12 | 2019-09-13 | 中国神华能源股份有限公司 | Repair method and system based on track car team state of the art |
| CN110717379A (en) * | 2019-08-28 | 2020-01-21 | 南京康尼机电股份有限公司 | Health assessment method for subway car door key components based on feature fusion |
-
2021
- 2021-05-28 CN CN202110594577.XA patent/CN113177650A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050060251A (en) * | 2003-12-16 | 2005-06-22 | 한국철도기술연구원 | Prevention system and the method for maintenance of railway car using rams analysis |
| CN109948169A (en) * | 2017-12-20 | 2019-06-28 | 中国中车股份有限公司 | A kind of railway freight-car prognostic and health management system |
| CN108520321A (en) * | 2018-04-03 | 2018-09-11 | 广西大学 | A kind of train bogie critical component preventative maintenance decision optimization model based on minimax ant group algorithm |
| CN109934358A (en) * | 2019-01-30 | 2019-06-25 | 中国人民解放军32181部队 | Equipment failure prediction and health evaluating method, system and terminal device |
| CN109886485A (en) * | 2019-02-13 | 2019-06-14 | 上海大学 | A kind of traffic infrastructure maintenance policy optimization method |
| CN110232452A (en) * | 2019-06-12 | 2019-09-13 | 中国神华能源股份有限公司 | Repair method and system based on track car team state of the art |
| CN110717379A (en) * | 2019-08-28 | 2020-01-21 | 南京康尼机电股份有限公司 | Health assessment method for subway car door key components based on feature fusion |
Non-Patent Citations (3)
| Title |
|---|
| 杨长风等: "《北斗卫星导航系统智能运行维护理论与实践（第1版）》", vol. 1, 31 October 2020, 中国宇航出版社, pages: 181 - 182 * |
| 罗亚萍: "汽车制动手柄预防维修周期优化模型研究", 《汽车实用技术》, no. 9, 15 May 2018 (2018-05-15), pages 134 - 136 * |
| 魏曙光等: "《人工智能在电气工程自动化的应用探索（第1版）》", vol. 1, 30 September 2020, 重庆大学出版社, pages: 166 - 167 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113988326A (en) * | 2021-10-09 | 2022-01-28 | 南京理工大学 | Subway equipment maintenance optimization method and system |
| CN113792944A (en) * | 2021-11-16 | 2021-12-14 | 深圳普菲特信息科技股份有限公司 | Predictive maintenance method and system |
| CN113792944B (en) * | 2021-11-16 | 2022-03-11 | 深圳普菲特信息科技股份有限公司 | Predictive maintenance method and system |
| CN114707918A (en) * | 2021-12-07 | 2022-07-05 | 西南交通大学 | Railway hierarchical node transportation method |
| CN114707918B (en) * | 2021-12-07 | 2023-04-07 | 西南交通大学 | Railway hierarchical node transportation method |
| CN115034306A (en) * | 2022-06-13 | 2022-09-09 | 昆明理工大学 | Aluminum electrolysis fault prediction and safe operation and maintenance method based on extensible neural network |
| CN117152930A (en) * | 2023-08-04 | 2023-12-01 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway cargo transportation state early warning method and device and electronic equipment |
| CN117152930B (en) * | 2023-08-04 | 2024-07-12 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway cargo transportation state early warning method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113177650A (en) | Predictive maintenance method and device for wagon compartment | |
| CN110764493B (en) | PHM application system, method and storage medium suitable for high-speed railway | |
| CN103617110B (en) | Server device condition maintenance system | |
| CN111401583A (en) | Escalator full life cycle health management system based on predictive maintenance | |
| CN111114519B (en) | Railway vehicle brake fault prediction method and health management system | |
| CN108629430A (en) | A kind of substantial equipment intelligence operation management system | |
| CN111489071A (en) | Maintenance method and system for rail transit vehicle | |
| CN111071291B (en) | Train wheel set monitoring system and train wheel set monitoring method | |
| CN106570567A (en) | Main network maintenance multi-constraint multi-target evaluation expert system and optimization method | |
| Brahimi et al. | Development of a prognostics and health management system for the railway infrastructure—Review and methodology | |
| Tichý et al. | Predictive diagnostics usage for telematic systems maintenance | |
| CN110262460B (en) | Concrete piston fault prediction method for extracting features by combining clustering idea | |
| Putra et al. | Designing machine learning model for predictive maintenance of railway vehicle | |
| Zvolenský et al. | Improved method of processing the output parameters of the diesel locomotive engine for more efficient maintenance | |
| KR101242060B1 (en) | The Reliability and Availability Management System of the Urban Transit EMU' Maintenance | |
| RU2579981C1 (en) | Automated systems for integrated management of resources, risk, reliability of railway transport | |
| RU2569216C2 (en) | Method of control over servicing and repair of railway rolling stock and system to this end | |
| Nappi | Integrated maintenance: analysis and perspective of innovation in railway sector | |
| CN117991742A (en) | Method, device and system for vehicle online diagnosis and operation and maintenance | |
| Attanayake et al. | On the necessity of using supervised machine learning for risk-based screening of distribution transformers: An industrial case study | |
| Najjar et al. | Predictive maintenance of Urban Metro Vehicles: Classification of air production unit failures using machine learning | |
| CN115034408A (en) | Internet of things maintenance prediction system for key equipment in seamless steel tube production | |
| Babishin et al. | An algorithm for estimating the effect of maintenance on aggregated covariates with application to railway switch point machines | |
| CN114066163A (en) | Power equipment state analysis method and device based on three-layer architecture mode | |
| CN118096131B (en) | Operation and maintenance inspection method based on electric power scene model |
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 |