CN113221382A - Method, system and equipment for predicting residual life of industrial equipment - Google Patents
Method, system and equipment for predicting residual life of industrial equipment Download PDFInfo
- Publication number
- CN113221382A CN113221382A CN202110610342.5A CN202110610342A CN113221382A CN 113221382 A CN113221382 A CN 113221382A CN 202110610342 A CN202110610342 A CN 202110610342A CN 113221382 A CN113221382 A CN 113221382A
- Authority
- CN
- China
- Prior art keywords
- domain
- data
- training
- life prediction
- residual life
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000012549 training Methods 0.000 claims abstract description 112
- 238000009826 distribution Methods 0.000 claims abstract description 41
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims description 36
- 238000012545 processing Methods 0.000 claims description 19
- 238000000605 extraction Methods 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 14
- 238000010606 normalization Methods 0.000 claims description 5
- 239000000126 substance Substances 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 8
- 238000006731 degradation reaction Methods 0.000 description 8
- 238000012360 testing method Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 230000015556 catabolic process Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000012546 transfer Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000011176 pooling Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012614 Monte-Carlo sampling Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013531 bayesian neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
Images
Classifications
-
- 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
-
- 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
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the field of prediction of residual life of industrial equipment, and particularly relates to a method, a system and equipment for predicting residual life of industrial equipment, aiming at solving the problems of difference of data distribution of different equipment in a characteristic space of a model and uncertainty of a prediction result. Acquiring data of industrial equipment with the end of service life as source domain data; acquiring data of running industrial equipment as target domain data; preprocessing source domain data and target domain data to obtain a source domain training data set and a target domain training data set; inputting a source domain training data set and a target domain training data set into a pre-constructed residual life prediction training model for training until the data distribution difference of two domains is minimized, and obtaining a residual life prediction model; and applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment. The invention greatly reduces the data distribution difference of the two domains and can also calculate the reliability of the residual life prediction result.
Description
Technical Field
The invention belongs to the technical field of prediction of residual life of industrial equipment, and particularly relates to a method, a system and equipment for predicting residual life of industrial equipment.
Background
In the actual industrial production, the degradation phenomenon of industrial equipment is difficult to avoid in the using process, and the real-time evaluation and prediction of the health state of the equipment are effective measures for reducing the probability of equipment failure and improving the safety guarantee.
Currently, there are three methods to predict the health status and failure rate of a device. The first is a physical modeling based approach, but it is difficult to meet the requirements of modern industrial systems in terms of workload and complexity. The second method is a data-driven method, which realizes the residual life prediction by analyzing the degradation data of the equipment operation process without depending on the physical characteristics of the equipment and the domain knowledge of the working principle, and has strong generalization performance, but the prediction accuracy is still left to be quotient discuss. And the third method is a deep learning-based method, and can complete modeling of the industrial equipment degradation process by means of deep feature extraction and complex nonlinear modeling capability, and achieve a good prediction effect.
However, since industrial equipment is interfered by external and internal uncertain factors in the degradation process, the deep learning-based method has two key problems:
1) the deep learning-based method can only realize point estimation of the residual service life of the industrial equipment generally, and can not measure the credibility of a prediction result.
2) Due to the fact that the data acquired by the target device and the data obtained during the training of the depth model are different under the interference of uncertain factors in the degradation process, different data distributions can be presented in the feature space of the model, the output result of the depth model is inaccurate, and the performance of the depth model is reduced in the application process.
Therefore, the method for predicting the residual service life of the industrial equipment can reduce the data distribution difference of different equipment, improve the generalization capability of the model, complete the uncertainty quantification of the residual service life prediction and output the credibility of the prediction result.
Disclosure of Invention
In order to solve the above problems in the prior art, namely, the problems of difference of data distribution of different equipment in a feature space of a model and uncertainty of a prediction result, the invention provides a method, a system and equipment for predicting the residual life of industrial equipment,
in a first aspect of the present invention, a method for predicting a remaining life of an industrial device is provided, where the method includes:
acquiring data of industrial equipment with the end of service life as source domain data; the source domain data comprises device full lifecycle data;
acquiring data of running industrial equipment as target domain data; the target domain data comprises device operational data;
preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the difference of data distribution of the two domains is minimized, and obtaining a residual life prediction model;
and applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
Optionally, the preprocessing comprises one or more processing operations of noise processing, missing value processing, normalization processing, wavelet transform, or fourier transform.
Optionally, the remaining life prediction training model includes a feature extraction module and a domain discrimination module, and the step of inputting the source domain training data set and the target domain training data set into the pre-constructed remaining life prediction training model for training until the difference between the data distribution of the two domains is minimized includes the following steps:
extracting a domain label characteristic value of a source domain training data set and a domain label characteristic value of a target domain training data set through the characteristic extraction module;
and inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module for training until the loss of the domain discrimination module is maximized, wherein the loss of the domain discrimination module is in inverse proportion to the data distribution difference of the two domains.
Optionally, the inputting the domain label feature values of the source domain and the target domain into the domain discriminant module for training until the loss of the domain discriminant module is maximized includes:
inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module to output a domain prediction result: the domain label characteristic value comprises a real domain label;
constructing a domain discrimination loss function according to the domain prediction result and the real domain label;
and adjusting parameters of a feature extraction module according to the loss value of the domain discriminant loss function until the loss value of the domain discriminant loss function is maximized.
Optionally, the remaining life prediction training model further includes a remaining life prediction module, and the method further includes:
extracting, by a feature extraction module, life-related feature values of the source domain training data set;
inputting the life-related characteristic value into the residual life prediction module for training until the loss of the residual life prediction module is minimized.
Optionally, the inputting the life-related feature value into the remaining life prediction module for training until the loss of the remaining life prediction module is minimized includes:
inputting the life-related characteristic value into the residual life prediction module to obtain a predicted life; the lifetime-related characteristic value comprises a true lifetime;
establishing a residual life prediction loss function according to the predicted life and the real life;
and adjusting parameters of the residual life prediction module according to the loss value of the residual life prediction loss function until the loss value of the residual life prediction loss function is minimized.
Optionally, the method further comprises:
and calculating the reliability of the residual life prediction result according to the predicted life and a reliability formula, wherein the reliability formula is as follows:
y is confidence, μtFor the average of the predicted remaining life of the T th of the T sample estimates,
the variance of the predicted remaining life for the T-th of the T sample estimates.
In a second aspect, the present invention provides a system for predicting remaining life of an industrial device, the system comprising
A first acquisition unit configured to acquire data of an industrial device whose lifetime is over as source domain data; the source domain data comprises device full lifecycle data;
a second acquisition unit for acquiring data of an operating industrial device as target domain data; the target domain data comprises device operational data;
the preprocessing unit is used for preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
the training unit is used for inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the data distribution difference of the two domains is minimized, and obtaining the residual life prediction model;
and the application unit is used for applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
In a third aspect of the present invention, an apparatus is provided, which includes:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the content of the first and second substances,
the memory stores instructions executable by the processor for performing the method of predicting remaining life of an industrial device according to any one of the first aspect.
In a fourth aspect of the present invention, a computer-readable storage medium is provided, where computer instructions are stored, and the computer instructions are used for being executed by the computer to implement the method for predicting the remaining life of the industrial equipment according to the first aspect.
The invention has the beneficial effects that: according to the invention, a residual life prediction training model is constructed and consists of a feature extraction module, a residual life prediction module and a field judgment module, the residual life prediction module can be trained through source domain data on one hand, so that the predicted life value can be infinitely close to the real life, and on the other hand, the field judgment module can be trained by combining target domain data, so that the field judgment module cannot distinguish whether the features extracted by the feature extraction module are from a target domain or a source domain, further more common features of the two domains are extracted, the data distribution difference of the two domains is reduced, the residual life prediction module is assisted to better transfer the existing knowledge of the source domain to the target domain for training, and the performance of the whole residual life prediction model is improved. In addition, the reliability of the residual life prediction result can be calculated.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
FIG. 1 is a schematic diagram of a method for predicting remaining life of an industrial device according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a method for predicting remaining life of an industrial device according to another embodiment of the present disclosure;
FIG. 3 is a schematic diagram of a residual life prediction training model according to an embodiment of the present application;
FIG. 4 is the predicted performance of the residual life prediction model on the training data bearing 1_ 1;
FIG. 5 is a predicted performance of the residual life prediction model on the test data bearing 1_ 3;
FIG. 6 is a schematic diagram of an industrial equipment remaining life prediction system according to an embodiment of the present application
FIG. 7 is a block diagram of a computer system of a server for implementing embodiments of the method, system, and apparatus of the present application.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings.
The invention provides a method for predicting the residual life of industrial equipment, which comprises the following steps:
acquiring data of industrial equipment with the end of service life as source domain data; the source domain data comprises device full lifecycle data;
acquiring data of running industrial equipment as target domain data; the target domain data comprises device operational data;
preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the difference of data distribution of the two domains is minimized, and obtaining a residual life prediction model;
and applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
In order to more clearly explain the method for predicting the remaining life of the industrial equipment, the steps in the embodiment of the present invention are described in detail below with reference to fig. 1.
The method for predicting the remaining life of the industrial equipment in the first embodiment of the invention comprises the following steps S101 to S105, wherein the following steps are described in detail:
step S101: acquiring data of industrial equipment with the end of service life as source domain data; the source domain data includes device full lifecycle data.
In this step, the full life cycle data refers to data of the whole process from the first use to the end of the life of the industrial equipment, and includes some data in the operation process, such as the rotation speed of the equipment, the load of the equipment, the actual service life and the like. The residual life prediction module can be trained through the source domain data, so that the predicted life value can be infinitely close to the real life, and the domain discrimination module can be trained by combining the target domain data to further extract more common features of the two domains by being incapable of discriminating whether the features extracted by the feature extraction module are from the target domain or the source domain, so that the data distribution difference of the two domains is reduced, the residual life prediction module is assisted to better transfer the existing knowledge of the source domain into the target domain for training, and the performance of the whole residual life prediction model is improved.
Step S102: acquiring data of running industrial equipment as target domain data; the target domain data includes device operational data.
In the embodiment of the application, the industrial device in operation is the industrial device to be tested, and in the prior art, generally, only the data of the industrial device with the end of life needs to be acquired for training, but when the model trained by using the data of the industrial device with the end of life is applied to other industrial devices to be tested, the result output by the model is inaccurate due to certain difference between device data. Therefore, the data of some running industrial equipment are acquired as target domain data, and the target domain data is also equivalent to test data to test and correct the parameter performance of the residual life prediction model. The specific means is that the data distribution difference of the two domains is reduced through a domain distinguishing module in the residual life prediction model, the auxiliary residual life prediction module can better transfer the existing knowledge of the source domain to the target domain for training, and the performance of the whole residual life prediction model is improved.
Step S103: and preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set.
Optionally, the preprocessing comprises one or more processing operations of noise processing, missing value processing, normalization processing, wavelet transform, or fourier transform.
The source domain data and the target domain data are preprocessed in the same way, in a specific example, the preprocessing comprises normalization processing and wavelet transformation processing, firstly, the acquired source domain data and the acquired target domain data are normalized, and the normalization processing formula is as follows:
wherein the content of the first and second substances,
the processed sample points are normalized for the source domain data set,
normalizing the processed sample points for the target domain data,
for the ith sample point in the source domain data,
is the ith sample point, x, in the target domain dataminRepresenting the minimum value, x, in the sample pointmaxRepresenting the maximum value among the sample points.
And further processing the normalized sample by adopting wavelet transform to obtain a two-dimensional representation of the sample. The wavelet transformation formula is as follows:
where a is the zoom factor, b is the pan position, ψ*Is the complex conjugate wavelet mother function, and t is the sampling period.
Obtaining a final source domain training data set after wavelet transformation
And a target domain training dataset
Wherein the content of the first and second substances,
is the real life data of the industrial equipment. n istIs the number of samples.
Step S104: and inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the difference of the data distribution of the two domains is minimized, and obtaining the residual life prediction model.
Optionally, as shown in fig. 2, the remaining life prediction training model includes a feature extraction module and a domain discrimination module, the remaining life prediction training model adopts a bayesian neural network model, and the inputting the source domain training data set and the target domain training data set into a pre-constructed remaining life prediction training model for training until the difference between the data distributions of the two domains is minimized includes the following steps,
step S201: and extracting the domain label characteristic value of the source domain training data set and the domain label characteristic value of the target domain training data set through the characteristic extraction module.
In this step, the feature extraction module consists of a series of convolutional layers and pooling layers, where convolutional layers operate as follows:
wherein, represents the convolution operation,
and
respectively representing the layer l-1 and layer l characteristics of the sample point,
and
respectively representing the convolution kernel and the offset,
in order to be a non-linear activation function,
the convolved sample points.
And sampling the obtained features by using a maximum pooling layer after the convolutional layer, so that the feature dimension of the original sample is reduced.
Step S202: and inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module for training until the loss of the domain discrimination module is maximized. Wherein, the loss of the domain discrimination module is inversely proportional to the difference of the data distribution of the two domains.
The loss of the domain judging module is inversely proportional to the data distribution difference of the two domains, and the greater the loss of the domain judging module is, the smaller the data distribution difference of the two domains is, so that the loss of the domain judging module is maximized to reduce the data distribution difference of the two domains, the domain judging module cannot distinguish which domain the features extracted by the feature extracting module come from, further more common features of the two domains are extracted, the data distribution difference of the two domains is reduced, the residual life predicting module is assisted to better transfer the existing knowledge of the source domain to the target domain for training, and the performance of the whole residual life predicting model is improved.
It should be noted that the fields referred to in this application are not industrial fields, but are distinguished fields for different data sources.
Specifically, the inputting the domain label feature values of the source domain and the target domain into the domain discrimination module for training until the loss of the domain discrimination module is maximized includes:
inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module to output a domain prediction result: the domain label characteristic value comprises a real domain label;
constructing a domain discrimination loss function according to the domain prediction result and the real domain label;
in one example, the discriminant loss function is as follows:
wherein x isiFor the samples in the source domain training dataset and the target domain training dataset, Gf(xi) As a feature extractor, diAs a real field tag, GdIs a domain discriminator.
In one example, it may be assumed that d of the source domaini1, d of the target Domaini＝0。
And adjusting parameters of a feature extraction module according to the loss value of the domain discriminant loss function until the loss value of the domain discriminant loss function is maximized.
Optionally, as shown in fig. 3, the remaining life prediction training model further includes a remaining life prediction module, and the method further includes:
extracting, by a feature extraction module, life-related feature values of the source domain training data set;
inputting the life-related characteristic value into the residual life prediction module for training until the loss of the residual life prediction module is minimized.
Optionally, the inputting the life-related feature value into the remaining life prediction module for training until the loss of the remaining life prediction module is minimized includes:
inputting the life-related characteristic value into the residual life prediction module to obtain a predicted life; the lifetime-related characteristic value comprises a true lifetime;
and establishing a residual life prediction loss function according to the predicted life and the real life.
In one example, an approximated distribution q (w | θ) may be constructed by variational inference to approximate a posterior distribution p (w | D)S) The posterior distribution is the source domain data DSAnd substituting the prior distribution p (w), wherein the prior distribution is the real life distribution obtained according to the historical data of the industrial equipment, and the posterior distribution is the data distribution for predicting the residual life.
The predicted loss of remaining life function is obtained by minimizing the KL divergence between the two probability distributions of the approximate distribution and the posterior distribution as follows:
some derivation steps are omitted in the middle, and the formula (7) is derived from the formula (5).
Where p (w) is a prior distribution,
and w is a parameter to be optimized of the model, and is the prediction loss of the residual life prediction module.
To further optimize the remaining life loss function, q (w | θ) is expressed as w ═ g (θ, ∈) by the reparameterization technique, where ∈ is the noise parameter sampled from a parameterless distribution q (∈), and will be
Is rewritten as
Using monte carlo sampling, a new remaining life loss function is obtained as follows:
let KL (q |. theta) | p (w) | N lambda | theta |2Further optimizing the loss:
while assuming that the source domain training data set obeys a Gaussian distribution, will
And expanding according to Gaussian distribution to obtain a final optimized loss function as follows:
wherein
Training samples in a dataset for a source domain
The predicted average value of the remaining life of the vehicle,
training samples in a dataset for a source domain
Predicted residual life variance, W ═ W, b]For the model parameters to be optimized, λ is a hyperparameter for adjusting the loss ratio, constant is a constant.
And adjusting parameters of the residual life prediction module according to the loss value of the residual life prediction loss function until the loss value of the residual life prediction loss function is minimized.
In another embodiment, the overall loss function of a remaining life prediction model may also be established according to equations (5) and (10) as follows:
wherein λ is a hyper-parameter for adjusting the loss ratio.
Thus, only the loss function of the formula (11) needs to be optimized, and the residual life prediction loss function and the domain discriminant loss function do not need to be separately optimized.
Step S105: and applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
Optionally, the method further comprises:
and calculating the reliability of the residual life prediction result according to the predicted life and a reliability formula, wherein the reliability formula is as follows:
y is confidence, μtFor the average of the predicted remaining life of the T th of the T sample estimates,
the variance of the predicted remaining life for the T-th of the T sample estimates.
In order to verify the method for predicting the residual life of the industrial equipment, which is provided by the invention, IEEE PHM 2012data challenge is selected[1]The bearing data are verified, the accelerated degradation experiment is carried out on the bearing under different working conditions, the full life cycle data of the bearing are obtained, and the training data are shown in table 1.
TABLE 1
The training data set adopts four groups of bearing data collected under the working condition 1 and the working condition 2, the sampling frequency is 25.6kHz, the sampling is carried out every 10s, the sampling time lasts for 0.1s, and each sample has 2570 sample points.
In the training stage, firstly, an available data set is constructed by using four groups of bearing data, data preprocessing operation is carried out on the constructed available data set, and a constructed residual life prediction training model is trained by using the preprocessed available data set; the performance of the trained residual life prediction model on the bearing 1_1 is shown in fig. 4; therefore, the residual life prediction model can effectively learn the degradation characteristics of the training data, construct an accurate residual life prediction model, quantify the uncertainty of the degradation process while accurately predicting, and output the credibility of the predicted residual life.
In the testing stage, the trained residual life prediction model is applied to the preprocessed test data, and the prediction result of the bearing 1_3 is shown in fig. 5. To compare the performance of the present invention with other methods on remaining life prediction performance, the error between the predicted remaining life RUL and the true life actriul is calculated and expressed as:
the results are shown in Table 2
TABLE 2 comparison of predicted results
The result shows that the residual life prediction model provided by the invention obtains a good prediction result on the bearing data of the test set; the credibility of the model prediction result can be obtained while providing more accurate residual life prediction, and more reasonable maintenance decision can be made.
References cited herein:
[1]Nectoux P,Gouriveau R,Medjaher K,et al.PRONOSTIA: An experimental platform for bearings accelerated degradation tests[C].IEEE International Conference on Prognostics and Health Management, PHM'12.,2012:1-8.
[2]Guo L,Li N,Jia F,et al.A recurrent neural network based health indicator for remaining useful life prediction of bearings[J].Neurocomputing,2017,240:98-109.
[3]Zhu J,Chen N,Peng W.Estimation of Bearing Remaining Useful Life Based on Multiscale Convolutional Neural Network[J].IEEE Transactions on Industrial Electronics,2019,66(4):3208-3216.
based on the same inventive concept, a second embodiment of the present invention provides a system for predicting remaining life of industrial equipment, as shown in fig. 6, the system comprising
A first acquisition unit 601 configured to acquire data of an industrial device whose lifetime has ended as source domain data; the source domain data comprises device full lifecycle data;
a second obtaining unit 602, configured to obtain data of an operating industrial device as target domain data; the target domain data comprises device operational data;
a preprocessing unit 603, configured to preprocess the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
a training unit 604, configured to input the source domain training data set and the target domain training data set into a pre-constructed remaining life prediction training model for training until a difference in data distribution between two domains is minimized, and obtain a remaining life prediction model;
an applying unit 605, configured to apply the residual life prediction model to a real scene of residual life prediction of the industrial equipment.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process and related description of the system described above may refer to the corresponding process in the foregoing method embodiments, and will not be described herein again.
It should be noted that, the system for predicting remaining life of industrial equipment provided in the foregoing embodiment is only illustrated by the division of the functional modules, and in practical applications, the functions may be allocated to different functional modules according to needs, that is, the modules or steps in the embodiment of the present invention are further decomposed or combined, for example, the modules in the foregoing embodiment may be combined into one module, or may be further split into multiple sub-modules, so as to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not to be construed as unduly limiting the present invention.
An apparatus of a third embodiment of the invention comprises:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the content of the first and second substances,
the memory stores instructions executable by the processor for performing steps of a method for predicting remaining life of an industrial device.
A computer-readable storage medium of a fourth embodiment of the present invention stores computer instructions for execution by the computer to implement the steps of the method for predicting remaining life of industrial equipment.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes and related descriptions of the storage device and the processing device described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
Reference is now made to FIG. 7, which illustrates a block diagram of a computer system of a server for implementing embodiments of the method, system, and apparatus of the present application. The server shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present application.
As shown in fig. 7, the computer system includes a Central Processing Unit (CPU)701, which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. In the RAM 703, various programs and data necessary for system operation are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An Input/Output (I/O) interface 705 is also connected to the bus 704.
The following components are connected to the I/O interface 705: an input portion 706 including a keyboard, a mouse, and the like; an output section 707 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and a speaker; a storage section 708 including a hard disk and the like; and a communication section 709 including a Network interface card such as a LAN (Local Area Network) card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. A drive 710 is also connected to the I/O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read out therefrom is mounted into the storage section 708 as necessary.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 709, and/or installed from the removable medium 711. The computer program, when executed by a Central Processing Unit (CPU)701, performs the above-described functions defined in the method of the present application. It should be noted that the computer readable medium mentioned above in the present application may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present application, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In this application, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present application may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The terms "first," "second," and the like are used for distinguishing between similar elements and not necessarily for describing or implying a particular order or sequence.
The terms "comprises," "comprising," or any other similar term are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
So far, the technical solutions of the present invention have been described in connection with the preferred embodiments shown in the accompanying drawings, but it is apparent that the scope of the present invention is not limited to these specific embodiments, as will be readily understood by those skilled in the art. Equivalent changes or substitutions of related technical features can be made by those skilled in the art without departing from the principle of the invention, and the technical scheme after the changes or substitutions can fall into the protection scope of the invention.
Claims (10)
1. A method for predicting remaining life of industrial equipment, the method comprising:
acquiring data of industrial equipment with the end of service life as source domain data; the source domain data comprises device full lifecycle data;
acquiring data of running industrial equipment as target domain data; the target domain data comprises device operational data;
preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the difference of data distribution of the two domains is minimized, and obtaining a residual life prediction model;
and applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
2. The method of claim 1, wherein the pre-processing comprises one or more of noise processing, missing value processing, normalization processing, wavelet transformation, or fourier transformation.
3. The method according to claim 1, wherein the residual life prediction training model comprises a feature extraction module and a domain discrimination module, and the step of inputting the source domain training data set and the target domain training data set into the pre-constructed residual life prediction training model for training until the difference of the data distribution of the two domains is minimized comprises the following steps:
extracting a domain label characteristic value of a source domain training data set and a domain label characteristic value of a target domain training data set through the characteristic extraction module;
and inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module for training until the loss of the domain discrimination module is maximized, wherein the loss of the domain discrimination module is in inverse proportion to the data distribution difference of the two domains.
4. The method of claim 3, wherein the inputting the domain label feature values of the source domain and the target domain into the domain discriminant module for training until the loss of the domain discriminant module is maximized comprises:
inputting the domain label characteristic values of the source domain and the target domain into the domain discrimination module to output a domain prediction result: the domain label characteristic value comprises a real domain label;
constructing a domain discrimination loss function according to the domain prediction result and the real domain label;
and adjusting parameters of a feature extraction module according to the loss value of the domain discriminant loss function until the loss value of the domain discriminant loss function is maximized.
5. The method of claim 3, wherein the remaining life prediction training model further comprises a remaining life prediction module, the method further comprising:
extracting a life-related characteristic value of the source domain training data set through a characteristic extraction module;
inputting the life-related characteristic value into the residual life prediction module for training until the loss of the residual life prediction module is minimized.
6. The method of claim 5, wherein the inputting the life-related feature values into the remaining life prediction module for training until the loss of the remaining life prediction module is minimized comprises:
inputting the life-related characteristic value into the residual life prediction module to obtain a predicted life; the lifetime-related characteristic value comprises a true lifetime;
establishing a residual life prediction loss function according to the predicted life and the real life;
and adjusting parameters of a residual life prediction module according to the loss value of the residual life prediction loss function until the loss value of the residual life prediction loss function is minimized.
7. The method of claim 1, further comprising:
and calculating the reliability of the residual life prediction result according to the predicted life and a reliability formula, wherein the reliability formula is as follows:
y is confidence, μtFor the mean of the predicted remaining life of the T th of the T sample estimates,
the variance of the predicted remaining life for the T-th of the T sample estimates.
8. An industrial equipment residual life prediction system, characterized in that the system comprises
A first acquisition unit configured to acquire data of an industrial device whose lifetime is over as source domain data; the source domain data comprises device full lifecycle data;
a second acquisition unit for acquiring data of an operating industrial device as target domain data; the target domain data comprises device operational data;
the preprocessing unit is used for preprocessing the source domain data and the target domain data to obtain a source domain training data set and a target domain training data set;
the training unit is used for inputting the source domain training data set and the target domain training data set into a pre-constructed residual life prediction training model for training until the data distribution difference of the two domains is minimized, and obtaining the residual life prediction model;
and the application unit is used for applying the residual life prediction model to a real scene of the residual life prediction of the industrial equipment.
9. An apparatus, comprising:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the content of the first and second substances,
the memory stores instructions executable by the processor for execution by the processor to implement the method of predicting remaining life of an industrial device of any one of claims 1-7.
10. A computer-readable storage medium storing computer instructions for execution by the computer to implement the method for predicting remaining life of industrial equipment according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110610342.5A CN113221382A (en) | 2021-06-01 | 2021-06-01 | Method, system and equipment for predicting residual life of industrial equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110610342.5A CN113221382A (en) | 2021-06-01 | 2021-06-01 | Method, system and equipment for predicting residual life of industrial equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113221382A true CN113221382A (en) | 2021-08-06 |
Family
ID=77082398
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110610342.5A Pending CN113221382A (en) | 2021-06-01 | 2021-06-01 | Method, system and equipment for predicting residual life of industrial equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113221382A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116069565A (en) * | 2023-03-15 | 2023-05-05 | 北京城建智控科技股份有限公司 | Method and device for replacing board card |
| CN116257972A (en) * | 2022-11-29 | 2023-06-13 | 元始智能科技(南通)有限公司 | Equipment state evaluation method and system based on field self-adaption and federal learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110362933A (en) * | 2019-07-17 | 2019-10-22 | 哈尔滨工业大学 | Aircraft auxiliary power plant life-span prediction method based on transfer learning |
| CN110555273A (en) * | 2019-09-05 | 2019-12-10 | 苏州大学 | bearing life prediction method based on hidden Markov model and transfer learning |
| CN110619346A (en) * | 2019-07-26 | 2019-12-27 | 浙江大学 | Modeling method of sea surface moving target model based on data fusion |
| CN111753416A (en) * | 2020-06-17 | 2020-10-09 | 重庆大学 | Lithium ion battery RUL prediction method based on two-stage Wiener process |
| CN112101220A (en) * | 2020-09-15 | 2020-12-18 | 哈尔滨理工大学 | Rolling bearing service life prediction method based on unsupervised model parameter migration |
| CN112241608A (en) * | 2020-10-13 | 2021-01-19 | 国网湖北省电力有限公司电力科学研究院 | Lithium battery life prediction method based on LSTM network and transfer learning |
| 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 |
-
2021
- 2021-06-01 CN CN202110610342.5A patent/CN113221382A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110362933A (en) * | 2019-07-17 | 2019-10-22 | 哈尔滨工业大学 | Aircraft auxiliary power plant life-span prediction method based on transfer learning |
| CN110619346A (en) * | 2019-07-26 | 2019-12-27 | 浙江大学 | Modeling method of sea surface moving target model based on data fusion |
| CN110555273A (en) * | 2019-09-05 | 2019-12-10 | 苏州大学 | bearing life prediction method based on hidden Markov model and transfer learning |
| CN111753416A (en) * | 2020-06-17 | 2020-10-09 | 重庆大学 | Lithium ion battery RUL prediction method based on two-stage Wiener process |
| CN112101220A (en) * | 2020-09-15 | 2020-12-18 | 哈尔滨理工大学 | Rolling bearing service life prediction method based on unsupervised model parameter migration |
| CN112241608A (en) * | 2020-10-13 | 2021-01-19 | 国网湖北省电力有限公司电力科学研究院 | Lithium battery life prediction method based on LSTM network and transfer learning |
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| HUANJIEWANG等: "Uncertainty Quantification of Bearing Remaining Useful Life Based on Convolutional Neural Network", 《2020 IEEE SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE (SSCI)》, 5 January 2021 (2021-01-05), pages 2893 - 2900 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116257972A (en) * | 2022-11-29 | 2023-06-13 | 元始智能科技(南通)有限公司 | Equipment state evaluation method and system based on field self-adaption and federal learning |
| CN116257972B (en) * | 2022-11-29 | 2024-02-20 | 元始智能科技(南通)有限公司 | Equipment state evaluation method and system based on field self-adaption and federal learning |
| CN116069565A (en) * | 2023-03-15 | 2023-05-05 | 北京城建智控科技股份有限公司 | Method and device for replacing board card |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112560876A (en) | Single-stage small sample target detection method for decoupling measurement | |
| CN110443222B (en) | Method and device for training face key point detection model | |
| CN113221382A (en) | Method, system and equipment for predicting residual life of industrial equipment | |
| CN115122155B (en) | Machine tool remote diagnosis method and system based on industrial internet big data | |
| CN111723728A (en) | Pedestrian searching method, system and device based on bidirectional interactive network | |
| CN113837596B (en) | Fault determination method and device, electronic equipment and storage medium | |
| CN112949767A (en) | Sample image increment, image detection model training and image detection method | |
| CN113158909A (en) | Behavior identification lightweight method, system and equipment based on multi-target tracking | |
| CN112800616A (en) | Equipment residual life self-adaptive prediction method based on proportional acceleration degradation modeling | |
| CN109088793B (en) | Method and apparatus for detecting network failure | |
| CN117593115A (en) | Feature value determining method, device, equipment and medium of credit risk assessment model | |
| Liu et al. | Software reliability forecasting: singular spectrum analysis and ARIMA hybrid model | |
| US20220148290A1 (en) | Method, device and computer storage medium for data analysis | |
| CN117290679A (en) | Running state detection method and device of current transformer and electronic equipment | |
| CN111259926A (en) | Meat freshness detection method and device, computing equipment and storage medium | |
| CN115293735A (en) | Unmanned factory industrial internet platform monitoring management method and system | |
| CN114782822A (en) | Method and device for detecting state of power equipment, electronic equipment and storage medium | |
| CN113297195A (en) | Time series abnormity detection method, device and equipment | |
| CN109978384A (en) | A kind of the leading factor analysis method and Related product of power distribution network operational efficiency | |
| CN117763329A (en) | Signal time-frequency characteristic extraction method and device based on iteration data | |
| CN111898080B (en) | Data sequence denoising method and device, electronic equipment and computer storage medium | |
| Grohmann et al. | On Learning Parametric Dependencies from Monitoring Data | |
| CN114266941A (en) | Method for rapidly detecting annotation result data of image sample | |
| CN115601839A (en) | Attendance checking method and device, electronic equipment and storage medium | |
| CN115830684A (en) | Personnel information detection method and system applied to industrial Internet and electronic equipment |
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 |