CN114444231A - Online self-adaptive prediction method, device, equipment and medium for residual life of mold - Google Patents
Online self-adaptive prediction method, device, equipment and medium for residual life of mold Download PDFInfo
- Publication number
- CN114444231A CN114444231A CN202210340203.XA CN202210340203A CN114444231A CN 114444231 A CN114444231 A CN 114444231A CN 202210340203 A CN202210340203 A CN 202210340203A CN 114444231 A CN114444231 A CN 114444231A
- Authority
- CN
- China
- Prior art keywords
- mold
- monitoring data
- residual life
- adaptive prediction
- online
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/12—Timing analysis or timing optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
- General Factory Administration (AREA)
Abstract
The application discloses a method, a device, equipment and a medium for online self-adaptive prediction of residual life of a mold, wherein the method for online self-adaptive prediction of the residual life of the mold comprises the following steps: acquiring multi-dimensional monitoring data and processing parameter information of the current operation state of a mold to be evaluated, performing feature extraction on the multi-dimensional monitoring data to obtain target feature information, performing online performance evaluation on the mold to be evaluated based on the target feature information to obtain a performance evaluation result, and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the processing parameter information to obtain a residual life prediction result of the mold to be evaluated, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different processing working condition states. The application solves the technical problem that the accuracy of the residual service life of the die is lower.
Description
Technical Field
The application relates to the technical field of Internet of things, in particular to a method, a device, equipment and a medium for online self-adaptive prediction of residual life of a mold.
Background
In recent years, with the advance of the internet of things of industry and the development of sensor technology, big data and artificial intelligence technology, the prediction and health management technology (PHM) is receiving more and more attention. The remaining life prediction technology is a key part in the PHM, and at present, the traditional mold remaining life prediction method usually predicts according to single collected signal data and does not consider the influence of different working modes on the remaining life, so that the accuracy of the remaining life of the mold is low.
Disclosure of Invention
The application mainly aims to provide a method, a device, equipment and a medium for online self-adaptive prediction of the residual life of a mold, and aims to solve the technical problem that the accuracy of the residual life of the mold in the prior art is low.
In order to achieve the above object, the present application provides an online adaptive prediction method for remaining life of a mold, including:
acquiring multi-dimensional monitoring data and processing parameter information of the current running state of a mold to be evaluated, wherein the multi-dimensional monitoring data comprises signal monitoring data;
performing feature extraction on the multidimensional monitoring data to obtain target feature information;
on the basis of the target characteristic information, carrying out online evaluation on the mold to be evaluated to obtain a performance evaluation result;
and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the machining parameter information to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different machining working condition states.
Optionally, the step of performing adaptive prediction by using a trained remaining life prediction model based on the performance evaluation result and the processing parameter information to obtain a remaining life prediction result of the mold to be evaluated includes:
based on a preset regression model, carrying out normalization processing on the processing parameter information to obtain a target working condition stress coefficient;
and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the target working condition stress coefficient to obtain a residual life prediction result of the to-be-evaluated mold.
Optionally, before the step of collecting the multidimensional monitoring data of the current operating state of the mold to be evaluated and the processing parameter information, the online adaptive prediction of the remaining life of the mold further includes:
acquiring state monitoring data of a mold to be trained under different processing working condition states and parameter information corresponding to the different working condition states, wherein the state monitoring data comprises data marked with a working condition label and a residual life value label;
training a regression model to be trained based on the parameter information of the different processing working condition states to obtain a preset regression model, and outputting stress coefficients of the different processing working condition states;
and performing iterative training on the prediction model to be trained on the basis of the stress coefficients of the different machining working condition states and the state monitoring data to obtain the residual life prediction model.
Optionally, the step of performing feature extraction on the multidimensional monitoring data to obtain target feature information includes:
filtering each signal monitoring data to obtain each target monitoring data;
extracting time series characteristics corresponding to the target monitoring data based on the timestamp labels corresponding to the target monitoring data;
respectively extracting time domain characteristics, frequency domain characteristics and time-frequency domain characteristics corresponding to each target monitoring data;
and performing feature fusion on the time sequence features, the time domain features, the frequency domain features and the time-frequency domain features to obtain the target feature information.
Optionally, the step of respectively extracting the time domain feature, the frequency domain feature and the time-frequency domain feature corresponding to each target monitoring data includes:
determining frequency domain signal data, time domain signal data and time-frequency domain signal data based on signal attributes corresponding to the target monitoring data;
extracting wavelet characteristic information corresponding to each time-frequency domain signal data, and taking the wavelet characteristic information as the time-frequency domain characteristics; calculating statistical characteristic information corresponding to each frequency domain signal data, and taking the statistical characteristic information as the time domain characteristic; and extracting frequency domain characteristics corresponding to the frequency domain signal data.
Optionally, the step of performing online evaluation on the mold to be evaluated based on the target feature information to obtain a performance evaluation result includes:
respectively judging the abnormality of the target characteristic information through a preset characteristic distribution model and a preset virtual measurement algorithm to obtain a first evaluation result;
evaluating the target characteristic information based on a pre-constructed data model to obtain a second evaluation result;
determining the performance evaluation result based on the first evaluation result and the second evaluation result.
Optionally, the step of collecting the multidimensional monitoring data of the mold to be evaluated in the current working condition state includes:
and acquiring multi-dimensional monitoring data of the mold to be evaluated in the current working condition state through preset sensors, wherein the multi-dimensional sensors comprise one or more of an acoustic emission sensor, an eddy current sensor, an ultrasonic sensor, a force sensor, a vibration sensor and a tonnage meter.
The application also provides an online self-adaptation of mould remaining life prediction device, the online self-adaptation of mould remaining life prediction device is virtual device, the online self-adaptation of mould remaining life prediction device includes:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring multi-dimensional monitoring data and processing parameter information of the current running state of a mold to be evaluated, and the multi-dimensional monitoring data comprises signal monitoring data;
the characteristic extraction module is used for extracting characteristics of the multi-dimensional monitoring data to obtain target characteristic information;
the performance evaluation module is used for carrying out online evaluation on the mold to be evaluated based on the target characteristic information to obtain a performance evaluation result;
and the self-adaptive prediction module is used for performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the machining parameter information to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different machining working condition states.
The application also provides online self-adaptive prediction equipment for the residual life of the die, which is entity equipment, and comprises: the online self-adaptive prediction method comprises a memory, a processor and an online self-adaptive prediction program of the residual service life of the die, wherein the online self-adaptive prediction program of the residual service life of the die is stored on the memory, and the online self-adaptive prediction program of the residual service life of the die is executed by the processor to realize the steps of the online self-adaptive prediction method of the residual service life of the die.
The application also provides a medium which is a computer readable medium, wherein the computer readable medium stores a mould residual life online adaptive prediction program, and the mould residual life online adaptive prediction program is executed by a processor to realize the steps of the mould residual life online adaptive prediction method.
The application provides an online self-adaptive prediction method, device, equipment and medium for the residual life of a mold, the method comprises the steps of firstly collecting multidimensional monitoring data and processing parameter information of the current operation state of the mold to be evaluated, wherein the multidimensional monitoring data comprise signal monitoring data, further carrying out feature extraction on the multidimensional monitoring data to obtain target feature information, further carrying out online evaluation on the mold to be evaluated based on the target feature information to obtain a performance evaluation result, further carrying out self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the processing parameter information to obtain a residual life prediction result of the mold to be evaluated, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different processing states, the problems of low precision and poor monitoring stability of single signal data are solved by acquiring multi-dimensional signal monitoring data, the flexibility and the monitoring accuracy are effectively improved, the multi-dimensional monitoring data are subjected to feature extraction, multi-dimensional feature information is mined, furthermore, the performance evaluation result of the mold is evaluated on line based on the feature information, the effective residual life of the mold is adaptively predicted by combining the performance evaluation result of the mold and the processing parameter information of a specific processing mode, and the accuracy of the residual life prediction of the mold is improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and together with the description, serve to explain the principles of the application.
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 described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a first embodiment of a method for online adaptive prediction of remaining life of a mold according to the present application;
FIG. 2 is a schematic flow chart of a second embodiment of the method for online adaptive prediction of remaining life of a mold according to the present application;
FIG. 3 is a schematic flow chart of a third embodiment of the method for online adaptive prediction of remaining life of a mold according to the present application;
FIG. 4 is a schematic flow chart of a fourth embodiment of the online adaptive prediction method for residual life of a mold according to the present application;
FIG. 5 is a schematic structural diagram of a device for online adaptive prediction of remaining life of a mold in a hardware operating environment according to an embodiment of the present application;
FIG. 6 is a schematic diagram of functional modules of the online self-adaptive residual life predicting device for the mold of the present application.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In a first embodiment of the online adaptive prediction method for the remaining life of the mold according to the present application, referring to fig. 1, the online adaptive prediction method for the remaining life of the mold includes:
step S10, collecting multidimensional monitoring data and processing parameter information of the current running state of the die to be evaluated, wherein the multidimensional monitoring data comprise signal monitoring data;
in this embodiment, it should be noted that the multidimensional monitoring data is signal data of the die in the die-casting process acquired on line by using various sensors, where the various sensors include, but are not limited to, acoustic emission sensors, eddy current sensors, ultrasonic sensors, force sensors, vibration sensors, tonnage meters, and the like, where the ultrasonic sensors: can provide various real-time data at die-casting in-process mould, in time monitor and prevent proruption unusual, acoustic emission sensor: mounting according to actual production requirements for collecting production state information, wherein the eddy current sensor comprises: capturing static and dynamic relative displacement changes between the metal conductor and the end face of the sensor, and vibrating the sensor: the vibration signal that is used for gathering in the course of working, the tonnage meter: the strain gauge is stretched and deformed by the strain of the machine body, so that the characteristics such as resistance and the like are changed, and the tonnage is further detected and converted. In addition, the processing parameter information of the processing working condition can be obtained through the controller signal.
Specifically, through preset sensors, signal monitoring data of the current operation state of the mold to be evaluated are acquired on line, the signal monitoring data are used as the multidimensional monitoring data, and in addition, processing parameter information corresponding to the current operation state, such as processing load, processing rate, process parameters and the like, is acquired.
Step S20, extracting the characteristics of the multidimensional monitoring data to obtain target characteristic information;
in this embodiment, specifically, the multidimensional monitoring data is filtered to obtain each target monitoring data, so as to improve the quality of signal data, and then time domain signal data, frequency domain signal data, and time-frequency domain signal data are extracted based on each target monitoring data, and then time domain features, frequency domain features, and time-frequency domain features are extracted based on the time domain signal data, the frequency domain signal data, and the time-frequency domain signal data.
Step S30, based on the target characteristic information, carrying out on-line evaluation on the mould to be evaluated to obtain a performance evaluation result;
in this embodiment, it should be noted that the online evaluation includes evaluation by mechanism information and evaluation by a data model.
Specifically, signal data of each sensor is collected in advance, and manually labeled based on field experience, signal features of the signal data in a healthy state are extracted, a distribution model is constructed, and further, the target feature information is input into the distribution model and compared with a preset confidence threshold value, so that a performance evaluation result predicted by the distribution model is determined based on a comparison result. And then, performing prediction evaluation on the signal monitoring data of the current running state through a preset virtual measurement technology to obtain a performance evaluation result corresponding to the virtual measurement technology, wherein the virtual measurement technology is used for solving the problem that the key quality of the machining process cannot be actually measured, the key index result of the produced product is estimated by utilizing the signal data of the production machine in the target characteristic information, the performance evaluation result of the mold is determined based on the key index result, and the performance evaluation result predicted by the distribution model and the performance evaluation result corresponding to the virtual measurement technology are used as the first evaluation result of the mechanism information.
Further, the target characteristic information is evaluated on line through a pre-constructed data model to obtain a second evaluation result, wherein the data model comprises models such as a principal component analysis algorithm, a logistic regression model, a self-organizing mapping network model and the like, in addition, the selection of a modeling method depends on an application scene, and besides the modeling method, other statistical methods and machine learning methods can be applied to signal data processing and modeling, so that the evaluation of the actual state performance of the current operation filling of the mold is realized.
And then determining the actual performance evaluation result of the current die to be evaluated based on the first evaluation result of the mechanism information and the second evaluation result of the data model, thereby realizing the combination of the mechanism information and the artificial intelligence model and improving the accuracy of the prediction of the residual life of the die.
And step S40, based on the performance evaluation result and the processing parameter information, performing self-adaptive prediction through a trained residual life prediction model to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different processing working condition states.
In this embodiment, it should be noted that the effective remaining life of the mold is related to the actual machining condition, and therefore, in this embodiment, the machining parameter information of the current operating state is taken into consideration when the remaining life is predicted. Specifically, the processing parameter information of the current operation state is input into a preset regression model, wherein the preset regression model is constructed based on the parameter information under different processing conditions, the processing parameter information is subjected to regression processing through the preset regression model to obtain a target output result, the target output result and the performance evaluation result are input into the residual life prediction model to realize self-adaptive model adjustment based on the target output result of the preset regression model to obtain the residual life prediction result of the to-be-evaluated mold, so that the use state of the model is evaluated through the preset regression model according to the difference of the processing parameter information such as processing load and process parameters, and the output of the model is used as an input parameter of the residual life prediction model, and the online self-adaptive mold life prediction is realized, and the performance evaluation result of the die and the processing parameter information of the specific processing state are combined to carry out self-adaptive prediction, so that the accuracy of the prediction of the effective residual life of the die is improved.
The embodiment of the application provides an online self-adaptive prediction method for the residual life of a mold, which comprises the steps of firstly collecting multidimensional monitoring data and processing parameter information of the current operation state of the mold to be evaluated, wherein the multidimensional monitoring data comprise signal monitoring data, further carrying out feature extraction on the multidimensional monitoring data to obtain target feature information, further carrying out online evaluation on the mold to be evaluated based on the target feature information to obtain a performance evaluation result, further carrying out self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the processing parameter information to obtain a residual life prediction result of the mold to be evaluated, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different processing conditions, the problems of low precision and poor monitoring stability of single signal data are solved by acquiring multi-dimensional signal monitoring data, the flexibility and the monitoring accuracy are effectively improved, the multi-dimensional monitoring data are subjected to feature extraction, multi-dimensional feature information is mined, furthermore, the performance evaluation result of the mold is evaluated on line based on the feature information, the effective residual life of the mold is adaptively predicted by combining the performance evaluation result of the mold and the processing parameter information of a specific processing mode, and the accuracy of the residual life prediction of the mold is improved.
Further, referring to fig. 2, based on the first embodiment in the present application, in another embodiment of the present application, step S20: performing feature extraction on the multidimensional monitoring data to obtain target feature information, which specifically comprises the following steps:
step S21, filtering each signal monitoring data to obtain each target monitoring data;
in this embodiment, it should be noted that, usually, the collected signal monitoring data includes noise information, and a filtering operation needs to be performed to filter each of the signal monitoring data, so as to improve the signal quality.
Step S22, extracting time series characteristics corresponding to each target monitoring data based on the timestamp label corresponding to each target monitoring data;
in this embodiment, it should be noted that the time-series characteristic includes signal monitoring data with a time stamp tag recorded in the whole die-casting manufacturing process, where the process data includes, but is not limited to, pressure time variation curve, temperature time variation curve, and flow time variation curve of the die-casting process. Specifically, feature extraction is performed according to the signal monitoring data and by combining process experience, so as to obtain time series feature information with time series attributes, for example, features such as duration of a pressure maintaining section, a signal section warping degree value, and temperature change gradients of different processing sections are extracted.
Step S23, respectively extracting time domain characteristics, frequency domain characteristics and time-frequency domain characteristics corresponding to each target monitoring data;
step S231, determining frequency domain signal data, time domain signal data, and time-frequency domain signal data based on the signal attribute corresponding to each of the target monitoring data;
step S232, extracting wavelet characteristic information corresponding to each time-frequency domain signal data, and taking the wavelet characteristic information as the time-frequency domain characteristics; calculating statistical characteristic information corresponding to each frequency domain signal data, and taking the statistical characteristic information as the time domain characteristic; and extracting frequency domain characteristics corresponding to the frequency domain signal data.
In this embodiment, it should be noted that the time domain features include, but are not limited to: mean, kurtosis, peaktop, root mean square, slope, kurtosis factor, variance, standard deviation, skewness, peak factor, form factor, impulse factor, and margin factor, including but not limited to: the system comprises signal energy, center of gravity frequency, frequency variance, mean square frequency, power spectrum mean, power spectrum peak, power spectrum variance, power spectrum kurtosis and power spectrum margin, wherein the time-frequency domain characteristic is the percentage of frequency band energy of wavelet packet decomposition.
And step S24, performing feature fusion on the time sequence features, the time domain features, the frequency domain features and the time-frequency domain features to obtain the target feature information.
In this embodiment, specifically, the time series characteristic, the time domain characteristic, the frequency domain characteristic, and the time-frequency domain characteristic are subjected to characteristic fusion, so as to obtain multi-dimensional characteristic information, and the multi-dimensional characteristic information is used as the target characteristic information.
In the embodiment of the application, through the above steps, that is, filtering each signal monitoring data to obtain each target monitoring data, further extracting time series characteristics corresponding to each target monitoring data based on the timestamp label corresponding to each target monitoring data, further extracting time domain characteristics, frequency domain characteristics and time-frequency domain characteristics corresponding to each target monitoring data respectively, and then the time sequence characteristics, the time domain characteristics, the frequency domain characteristics and the time-frequency domain characteristics are subjected to characteristic fusion to obtain the target characteristic information, so that the fusion of effective characteristic information of different dimensions in multi-signal monitoring data is realized, the diversity and completeness of information extraction are ensured, the problems of low precision and poor monitoring stability of single signal data are solved, the flexibility and the monitoring accuracy are effectively improved, and the accuracy of residual life prediction of the die in the die-casting process is further improved.
Further, referring to fig. 3, based on the first embodiment in the present application, in another embodiment of the present application, step S40: based on the performance evaluation result and the processing parameter information, performing adaptive prediction through a trained residual life prediction model to obtain a residual life prediction result of the to-be-evaluated mold, specifically comprising:
step S41, based on a preset regression model, carrying out normalization processing on the processing parameter information to obtain a target working condition stress coefficient;
and step S42, based on the performance evaluation result and the target working condition stress coefficient, performing self-adaptive prediction through a trained residual life prediction model to obtain a residual life prediction result of the to-be-evaluated mold.
In this embodiment, it should be noted that, in the case that the parameters of the conventional offline model are generally fixed, and the conditions are the same or relatively similar, the accuracy of model prediction is relatively high, however, in the actual production process, the conditions of the die-casting process are varied, and therefore, the accuracy of the residual life prediction performed by using the offline model with the fixed parameters is relatively low.
It should be further noted that the preset regression model is constructed based on parameter information corresponding to different processing conditions, where the parameter information includes information such as processing load, processing rate, and process parameters, and further, the output of the preset regression model is used as an input parameter of the remaining life prediction model, and the remaining life prediction model is obtained by training in combination with the collected state monitoring data of different processing conditions.
Specifically, the processing parameter information is input into the preset regression model to be subjected to normalization and quantization processing, so that a target working condition stress coefficient is obtained, further, the target working condition stress coefficient and a performance evaluation result are input into a residual life prediction model, a residual life prediction result of the to-be-evaluated mold in the current operation state is obtained, and the accuracy of the residual life is realized.
Through the steps, namely, based on the preset regression model, the processing parameter information is subjected to normalization processing to obtain the target working condition stress coefficient, and then based on the performance evaluation result and the target working condition stress coefficient, the trained residual life prediction model is subjected to self-adaptive prediction to obtain the residual life prediction result of the to-be-evaluated mold, so that the processing parameter information of the current running state of the preset regression model is used for evaluating the service state of the model according to the target working condition stress coefficient of the current state, the self-adaptation of the residual life prediction model is realized, further, the effective residual life of the self-adaptive prediction mold is obtained by combining the mold performance evaluation result and the specific processing state, and the accuracy of model prediction is improved.
Further, referring to fig. 4, based on the first embodiment in the present application, in another embodiment of the present application, before step S10, the method further includes:
step A10, collecting state monitoring data of a die to be trained under different processing working conditions and parameter information corresponding to different working conditions, wherein the state monitoring data comprises data marked with a working condition label and a residual life value label;
step A20, training a regression model to be trained based on the parameter information of different processing working condition states to obtain the preset regression model, and outputting stress coefficients of different processing working condition states;
and A30, performing iterative training on the prediction model to be trained based on the stress coefficients of different machining working condition states and the state monitoring data to obtain the residual life prediction model.
In this embodiment, it should be noted that the state monitoring data in the different working condition states is the state monitoring data of the die in the whole life cycle of the die-casting process. Specifically, state monitoring data of a mold to be trained under different processing working condition states are acquired through a plurality of preset sensors, and parameter information corresponding to the different working condition states is acquired, wherein the state monitoring data comprise data marked with working condition labels and residual life value labels, the preset regression model is constructed based on the parameter information corresponding to the different working condition states, stress coefficients of the different processing working condition states are output, further, feature extraction is carried out on the state monitoring data based on the state monitoring data of the different processing working condition states, feature information of the different processing working condition states is acquired, further, iterative training is carried out on a prediction model to be trained based on the stress coefficients, the feature information and the residual life value labels of the different processing working condition states, and optimal parameter models corresponding to the different processing working condition states are acquired, and using the optimal parameter model corresponding to different processing working condition states as the residual life prediction model, and additionally, in an implementable mode, clustering the characteristic information belonging to the same working condition label based on the working condition label to obtain clustering characteristic information, so as to perform modeling according to the stress coefficient, the clustering characteristic information and the residual life value label of different processing working condition states, thereby enabling the stress coefficient of the current processing parameter information to be determined through a preset regression model through multi-dimensional monitoring data and processing parameter information of the current operating state acquired on line in the subsequent on-line prediction process, and further performing an adaptive adjustment model based on the stress coefficient to realize the adaptive residual life prediction based on the working condition.
Through the steps, namely, collecting the state monitoring data of the die to be trained under different processing working condition states and the parameter information corresponding to the different working condition states, wherein the state monitoring data comprises the data marked with the working condition labels and the residual life value labels, training the regression model to be trained based on the parameter information of the different processing working condition states to obtain the preset regression model, outputting the stress coefficients of the different processing working condition states, further, carrying out iterative training on the prediction model to be trained based on the stress coefficients of the different processing working condition states and the state monitoring data to obtain the residual life prediction model, realizing that the parameter information corresponding to the different working condition states constructs the preset regression model, and constructing the state monitoring data of the different working condition states and the stress coefficients of the different working condition states output by the preset regression model, and training to obtain optimal parameter models of different processing working condition states, thereby realizing the self-adaptive residual life prediction of the working conditions and improving the accuracy of the residual life prediction of the die.
Referring to fig. 5, fig. 5 is a schematic structural diagram of a device for online adaptive prediction of remaining life of a mold in a hardware operating environment according to an embodiment of the present application.
As shown in fig. 5, the online adaptive residual life prediction device for the mold may include: a processor 1001, such as a CPU, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to realize connection and communication between the processor 1001 and the memory 1005. The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 1005 may alternatively be a memory device separate from the processor 1001 described above.
Optionally, the online adaptive prediction device for remaining life of the mold may further include a rectangular user interface, a network interface, a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and the like. The rectangular user interface may comprise a Display screen (Display), an input sub-module such as a Keyboard (Keyboard), and the optional rectangular user interface may also comprise a standard wired interface, a wireless interface. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WIFI interface).
Those skilled in the art will appreciate that the die remaining life online adaptive prediction device configuration shown in FIG. 5 does not constitute a definition of a die remaining life online adaptive prediction device, and may include more or fewer components than those shown, or some components in combination, or a different arrangement of components.
As shown in fig. 5, a memory 1005, which is a kind of computer medium, may include therein an operating device, a network communication module, and a mold remaining life online adaptive prediction program. The operation device is a program for managing and controlling hardware and software resources of the online self-adaptive prediction device for the residual life of the mold, and supports the operation of the online self-adaptive prediction program for the residual life of the mold and other software and/or programs. The network communication module is used for realizing communication among the components in the memory 1005 and communication with other hardware and software in the online adaptive prediction device of the residual service life of the mold.
In the online adaptive prediction device for residual mold life shown in fig. 5, the processor 1001 is configured to execute an online adaptive prediction program for residual mold life stored in the memory 1005, and implement the steps of the online adaptive prediction method for residual mold life described in any one of the above.
The specific implementation of the online self-adaptive residual life predicting device for the mold is basically the same as that of each embodiment of the online self-adaptive residual life predicting method for the mold, and is not described herein again.
In addition, referring to fig. 6, fig. 6 is a schematic diagram of functional modules of the online adaptive predicting device for residual life of the mold according to the present application, and the present application further provides an online adaptive predicting device for residual life of the mold, which includes:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring multi-dimensional monitoring data and processing parameter information of the current running state of a mold to be evaluated, and the multi-dimensional monitoring data comprises signal monitoring data;
the characteristic extraction module is used for extracting characteristics of the multi-dimensional monitoring data to obtain target characteristic information;
the performance evaluation module is used for carrying out online evaluation on the mold to be evaluated based on the target characteristic information to obtain a performance evaluation result;
and the self-adaptive prediction module is used for performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the machining parameter information to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different machining working condition states.
Optionally, the adaptive prediction module is further configured to:
based on a preset regression model, carrying out normalization processing on the processing parameter information to obtain a target working condition stress coefficient;
and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the target working condition stress coefficient to obtain a residual life prediction result of the to-be-evaluated mold.
Optionally, the online adaptive residual life predicting device of the mold is further configured to:
acquiring state monitoring data of a mold to be trained under different processing working condition states and parameter information corresponding to the different working condition states, wherein the state monitoring data comprises data marked with a working condition label and a residual life value label;
training a regression model to be trained based on the parameter information of the different processing working condition states to obtain a preset regression model, and outputting stress coefficients of the different processing working condition states;
and performing iterative training on the prediction model to be trained on the basis of the stress coefficients of the different machining working condition states and the state monitoring data to obtain the residual life prediction model.
Optionally, the feature extraction module is further configured to:
filtering each signal monitoring data to obtain each target monitoring data;
extracting time series characteristics corresponding to the target monitoring data based on the timestamp labels corresponding to the target monitoring data;
respectively extracting time domain characteristics, frequency domain characteristics and time-frequency domain characteristics corresponding to each target monitoring data;
and performing feature fusion on the time sequence features, the time domain features, the frequency domain features and the time-frequency domain features to obtain the target feature information.
Optionally, the feature extraction module is further configured to:
determining frequency domain signal data, time domain signal data and time-frequency domain signal data based on signal attributes corresponding to the target monitoring data;
extracting wavelet characteristic information corresponding to each time-frequency domain signal data, and taking the wavelet characteristic information as the time-frequency domain characteristics; calculating statistical characteristic information corresponding to each frequency domain signal data, and taking the statistical characteristic information as the time domain characteristic; and extracting frequency domain characteristics corresponding to the frequency domain signal data.
Optionally, the performance evaluation module is further configured to:
respectively carrying out abnormity judgment on the multi-dimensional state feature information through a preset feature distribution model and a preset virtual measurement algorithm to obtain a first evaluation result;
evaluating the multi-dimensional state characteristic information based on a pre-constructed data model to obtain a second evaluation result;
determining the performance evaluation result based on the first evaluation result and the second evaluation result.
Optionally, the acquisition module is further configured to:
and acquiring multi-dimensional monitoring data of the mold to be evaluated in the current working condition state through preset sensors, wherein the multi-dimensional sensors comprise one or more of an acoustic emission sensor, an eddy current sensor, an ultrasonic sensor, a force sensor, a vibration sensor and a tonnage meter.
The specific implementation of the online adaptive prediction device for the residual life of the mold is basically the same as that of each embodiment of the online adaptive prediction method for the residual life of the mold, and is not described herein again.
The present application provides a medium, which is a computer readable medium, and the computer readable medium stores one or more programs, which can also be executed by one or more processors, for implementing the steps of the online adaptive prediction method for remaining life of a mold according to any one of the above methods.
The specific implementation manner of the computer readable medium of the present application is substantially the same as that of each embodiment of the above online adaptive prediction method for the remaining life of the mold, and is not described herein again.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (10)
1. The online self-adaptive prediction method for the residual life of the mold is characterized by comprising the following steps of:
acquiring multi-dimensional monitoring data and processing parameter information of the current running state of a mold to be evaluated, wherein the multi-dimensional monitoring data comprises signal monitoring data;
performing feature extraction on the multidimensional monitoring data to obtain target feature information;
on the basis of the target characteristic information, carrying out on-line evaluation on the mold to be evaluated to obtain a performance evaluation result;
and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the machining parameter information to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different machining working condition states.
2. The online adaptive prediction method for the residual life of the mold according to claim 1, wherein the step of performing adaptive prediction by using a trained residual life prediction model based on the performance evaluation result and the processing parameter information to obtain the residual life prediction result of the mold to be evaluated comprises:
based on a preset regression model, carrying out normalization processing on the processing parameter information to obtain a target working condition stress coefficient;
and performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the target working condition stress coefficient to obtain a residual life prediction result of the to-be-evaluated mold.
3. The online adaptive prediction method for the residual life of the mold according to claim 2, wherein before the step of collecting the multidimensional monitoring data of the current operating state of the mold to be evaluated and the processing parameter information, the online adaptive prediction method for the residual life of the mold further comprises:
acquiring state monitoring data of a mold to be trained under different processing working condition states and parameter information corresponding to the different working condition states, wherein the state monitoring data comprises data marked with a working condition label and a residual life value label;
training a regression model to be trained based on the parameter information of the different processing working condition states to obtain a preset regression model, and outputting stress coefficients of the different processing working condition states;
and performing iterative training on the prediction model to be trained on the basis of the stress coefficients of the different machining working condition states and the state monitoring data to obtain the residual life prediction model.
4. The online adaptive prediction method for the residual life of the mold according to claim 1, wherein the step of performing feature extraction on the multidimensional monitoring data to obtain target feature information comprises:
filtering each signal monitoring data to obtain each target monitoring data;
extracting time series characteristics corresponding to the target monitoring data based on the timestamp labels corresponding to the target monitoring data;
respectively extracting time domain characteristics, frequency domain characteristics and time-frequency domain characteristics corresponding to each target monitoring data;
and performing feature fusion on the time sequence features, the time domain features, the frequency domain features and the time-frequency domain features to obtain the target feature information.
5. The online adaptive prediction method for the residual life of the mold according to claim 4, wherein the step of respectively extracting the time domain feature, the frequency domain feature and the time-frequency domain feature corresponding to each target monitoring data comprises:
determining frequency domain signal data, time domain signal data and time-frequency domain signal data based on signal attributes corresponding to the target monitoring data;
extracting wavelet characteristic information corresponding to each time-frequency domain signal data, and taking the wavelet characteristic information as the time-frequency domain characteristics; calculating statistical characteristic information corresponding to each frequency domain signal data, and taking the statistical characteristic information as the time domain characteristic; and extracting frequency domain characteristics corresponding to the frequency domain signal data.
6. The online adaptive prediction method for the remaining life of the mold according to claim 1, wherein the online evaluation of the mold to be evaluated based on the target feature information to obtain the performance evaluation result comprises:
respectively judging the abnormality of the target characteristic information through a preset characteristic distribution model and a preset virtual measurement algorithm to obtain a first evaluation result;
evaluating the target characteristic information based on a pre-constructed data model to obtain a second evaluation result;
determining the performance evaluation result based on the first evaluation result and the second evaluation result.
7. The online adaptive prediction method for the residual life of the mold according to claim 1, wherein the step of collecting the multidimensional monitoring data of the mold to be evaluated in the current operating state comprises:
and acquiring multi-dimensional monitoring data of the mold to be evaluated in the current running state through preset sensors, wherein each sensor comprises one or more of an acoustic emission sensor, an eddy current sensor, an ultrasonic sensor, a force sensor, a vibration sensor and a tonnage meter.
8. The online self-adaptive prediction device for the residual life of the mold is characterized by comprising the following components:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring multi-dimensional monitoring data and processing parameter information of the current running state of a mold to be evaluated, and the multi-dimensional monitoring data comprises signal monitoring data;
the characteristic extraction module is used for extracting characteristics of the multi-dimensional monitoring data to obtain target characteristic information;
the performance evaluation module is used for carrying out online evaluation on the mold to be evaluated based on the target characteristic information to obtain a performance evaluation result;
and the self-adaptive prediction module is used for performing self-adaptive prediction through a trained residual life prediction model based on the performance evaluation result and the machining parameter information to obtain a residual life prediction result of the to-be-evaluated mold, wherein the residual life prediction model is obtained by training based on pre-collected state monitoring data under different machining working condition states.
9. The online adaptive prediction device for the residual life of the mold is characterized by comprising: a memory, a processor and a mould residual life online self-adaptive prediction program stored on the memory,
the online adaptive prediction program for residual life of the mold is executed by the processor to realize the online adaptive prediction method for residual life of the mold according to any one of claims 1 to 7.
10. A medium, which is a computer readable medium, wherein the computer readable medium has stored thereon an online adaptive prediction program for remaining life of a mold, the online adaptive prediction program for remaining life of a mold is executed by a processor to implement the steps of the online adaptive prediction method for remaining life of a mold according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210340203.XA CN114444231B (en) | 2022-04-02 | 2022-04-02 | Online self-adaptive prediction method, device, equipment and medium for residual life of mold |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210340203.XA CN114444231B (en) | 2022-04-02 | 2022-04-02 | Online self-adaptive prediction method, device, equipment and medium for residual life of mold |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114444231A true CN114444231A (en) | 2022-05-06 |
| CN114444231B CN114444231B (en) | 2022-07-12 |
Family
ID=81359059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210340203.XA Active CN114444231B (en) | 2022-04-02 | 2022-04-02 | Online self-adaptive prediction method, device, equipment and medium for residual life of mold |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114444231B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114749499A (en) * | 2022-06-13 | 2022-07-15 | 深圳市信润富联数字科技有限公司 | Descaling nozzle control method and device, electronic equipment and readable storage medium |
| CN116228028A (en) * | 2023-03-13 | 2023-06-06 | 平湖新仕鑫新材料股份有限公司 | Application performance evaluation method and system for plastic bags |
| CN116738859A (en) * | 2023-06-30 | 2023-09-12 | 常州润来科技有限公司 | Online nondestructive life assessment method and system for copper pipe |
| CN117172756A (en) * | 2023-11-01 | 2023-12-05 | 深圳市信润富联数字科技有限公司 | Mold management method, apparatus, device and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080208487A1 (en) * | 2007-02-23 | 2008-08-28 | General Electric Company | System and method for equipment remaining life estimation |
| US8781982B1 (en) * | 2011-09-23 | 2014-07-15 | Lockheed Martin Corporation | System and method for estimating remaining useful life |
| CN109948860A (en) * | 2019-03-26 | 2019-06-28 | 哈工大机器人（合肥）国际创新研究院 | A kind of mechanical system method for predicting residual useful life and system |
| CN111274737A (en) * | 2020-02-25 | 2020-06-12 | 山东大学 | Method and system for predicting remaining service life of mechanical equipment |
| CN111967133A (en) * | 2020-07-10 | 2020-11-20 | 西北工业大学 | Method for predicting residual life of cutter in multiple cutting stages |
| CN112765890A (en) * | 2021-01-26 | 2021-05-07 | 西安电子科技大学 | Dynamic domain adaptive network-based multi-working-condition rotating machine residual life prediction method |
| CN113467375A (en) * | 2021-01-06 | 2021-10-01 | 南京航空航天大学 | Method for predicting remaining usable life of intelligent numerical control machining tool based on virtual measurement |
| CN113723016A (en) * | 2021-11-01 | 2021-11-30 | 深圳市信润富联数字科技有限公司 | Punch residual life prediction method, device and system and readable storage medium |
-
2022
- 2022-04-02 CN CN202210340203.XA patent/CN114444231B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080208487A1 (en) * | 2007-02-23 | 2008-08-28 | General Electric Company | System and method for equipment remaining life estimation |
| US8781982B1 (en) * | 2011-09-23 | 2014-07-15 | Lockheed Martin Corporation | System and method for estimating remaining useful life |
| CN109948860A (en) * | 2019-03-26 | 2019-06-28 | 哈工大机器人（合肥）国际创新研究院 | A kind of mechanical system method for predicting residual useful life and system |
| CN111274737A (en) * | 2020-02-25 | 2020-06-12 | 山东大学 | Method and system for predicting remaining service life of mechanical equipment |
| CN111967133A (en) * | 2020-07-10 | 2020-11-20 | 西北工业大学 | Method for predicting residual life of cutter in multiple cutting stages |
| CN113467375A (en) * | 2021-01-06 | 2021-10-01 | 南京航空航天大学 | Method for predicting remaining usable life of intelligent numerical control machining tool based on virtual measurement |
| CN112765890A (en) * | 2021-01-26 | 2021-05-07 | 西安电子科技大学 | Dynamic domain adaptive network-based multi-working-condition rotating machine residual life prediction method |
| CN113723016A (en) * | 2021-11-01 | 2021-11-30 | 深圳市信润富联数字科技有限公司 | Punch residual life prediction method, device and system and readable storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 刘紫阳 等: "一种新的基于工况区分和轨迹相似性的RUL预测方法", 《电子测量技术》 * |
| 李琪等: "变工况条件下基于相似性的剩余使用寿命预测方法", 《北京航空航天大学学报》 * |
| 王奉涛 等: "基于KPCA和WPHM的滚动轴承可靠性评估与寿命预测", 《振动、测试与诊断》 * |
| 韩冰洁 等: "基于工况聚类分析和信息融合...相似性剩余使用寿命预测方法", 《探测与控制学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114749499A (en) * | 2022-06-13 | 2022-07-15 | 深圳市信润富联数字科技有限公司 | Descaling nozzle control method and device, electronic equipment and readable storage medium |
| CN116228028A (en) * | 2023-03-13 | 2023-06-06 | 平湖新仕鑫新材料股份有限公司 | Application performance evaluation method and system for plastic bags |
| CN116228028B (en) * | 2023-03-13 | 2023-12-19 | 平湖新仕鑫新材料股份有限公司 | Application performance evaluation method and system for plastic bags |
| CN116738859A (en) * | 2023-06-30 | 2023-09-12 | 常州润来科技有限公司 | Online nondestructive life assessment method and system for copper pipe |
| CN116738859B (en) * | 2023-06-30 | 2024-02-02 | 常州润来科技有限公司 | Online nondestructive life assessment method and system for copper pipe |
| CN117172756A (en) * | 2023-11-01 | 2023-12-05 | 深圳市信润富联数字科技有限公司 | Mold management method, apparatus, device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114444231B (en) | 2022-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114444231B (en) | Online self-adaptive prediction method, device, equipment and medium for residual life of mold | |
| CN113792453B (en) | Digital twinning-based partial discharge monitoring system, method and device | |
| JP7007243B2 (en) | Anomaly detection system | |
| JP2020500420A (en) | Yield prediction system and method for machine learning based semiconductor manufacturing | |
| WO2017008451A1 (en) | Abnormal load detecting method for cloud computing oriented online service | |
| CN106228270B (en) | Energy consumption prediction method and system for big data driven extrusion equipment | |
| CN113255777B (en) | Equipment fault early warning method and system based on multi-mode sensitive feature selection fusion | |
| CN109743356B (en) | Industrial internet data acquisition method and device, readable storage medium and terminal | |
| CN117274259B (en) | Keyboard production equipment management method and system based on Internet of things | |
| CN112002114A (en) | Electromechanical equipment wireless data acquisition system and method based on 5G-ZigBee communication | |
| CN117421994A (en) | Edge application health monitoring method and system | |
| CN114202256B (en) | Architecture upgrading early warning method and device, intelligent terminal and readable storage medium | |
| CN117891223B (en) | Equipment management method, system, equipment and medium based on industrial Internet of things | |
| CN113110961B (en) | Equipment abnormality detection method and device, computer equipment and readable storage medium | |
| JP7028133B2 (en) | Control system and control method | |
| CN114238383A (en) | Big data extraction method and device for supply chain monitoring | |
| CN117634905A (en) | Power grid business data feature extraction method | |
| Qing-Fang et al. | Small-time scale network traffic prediction based on a local support vector machine regression model | |
| CN111222203B (en) | Bearing service life model creation and prediction method thereof | |
| CN117272022A (en) | Detection method of MEMS oscillator | |
| US20190155269A1 (en) | Monitoring system and monitoring method | |
| CN115203905A (en) | Equipment health assessment method integrating expert experience and intelligent algorithm | |
| CN114624791A (en) | Rainfall measurement method and device, computer equipment and storage medium | |
| CN113986704A (en) | TS-Decomposition-based data center high-frequency fault time domain early warning method and system | |
| CN110275809B (en) | Data fluctuation identification method and device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |