CN108710958A - A kind of prediction health control method and device, computer readable storage medium - Google Patents
A kind of prediction health control method and device, computer readable storage medium Download PDFInfo
- Publication number
- CN108710958A CN108710958A CN201810469967.2A CN201810469967A CN108710958A CN 108710958 A CN108710958 A CN 108710958A CN 201810469967 A CN201810469967 A CN 201810469967A CN 108710958 A CN108710958 A CN 108710958A
- Authority
- CN
- China
- Prior art keywords
- layer
- data set
- neuron
- carried out
- bias
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
This application discloses a kind of prediction health control PHM method and devices, computer readable storage medium, the method includes：The data of sensor acquisition are obtained, chronologically adding window is divided into multiple data blocks by the data of acquisition；Concurrent Feature extraction is carried out to the multiple data blocks marked off by multithreading, and the feature category that each thread extracts is merged, the characteristic data set after being merged；Health evaluating, fault diagnosis or predicting residual useful life are carried out using obtained characteristic data set.The application is divided into multiple data blocks by the data time sequence windowing process for acquiring sensor, then Concurrent Feature extraction is carried out, provide a kind of health control platform for supporting concurrent operation, the upper PHM system operations speed of equipment and information analysis ability are improved, has effectively ensured the safe operation of equipment.
Description
Technical field
The present invention relates to field of computer technology, and in particular to a kind of prediction health control method and device, computer can
Read storage medium.
Background technology
Prediction health control (Prognostic and Health Management, PHM) is to utilize all kinds of advanced sensings
Device monitors equipment running state parameter and characteristic signal in real time, assesses equipment health status by intelligent algorithm and model, in advance
Remaining life is surveyed, the system for diagnosing fault type and a series of breakdown maintenance decision being provided before failure occurs.PHM skills
Art is the product being combined by advanced diagnostic techniques, measuring technology management of equipment maintenance theory.Equip user of service and repair
Personnel identify fault type by the trouble diagnosibility of this technology, to take effective maintenance mode, effectively reduce
Failure risk saves equipment resource, reduces the economic loss brought by equipment failure mistaken diagnosis.
But in some fields, the system structure of equipment is typically more complicated, part grade, component-level and the complete machine being related to
Grade key structure position is very more, various so as to cause the damage and failure complexity of equipment configuration, and diagnosis in the process also can be by
The influence of uncertain factor, therefore, it is necessary to multiple positions to equipment to be monitored in real time, the safe operation of Support Equipment.It is existing
There are the equipment PHM systems that technology is realized often to be divided into ground PHM systems and the upper PHM system two parts of equipment, but equips upper equipment
Signal processing and information analysis ability are limited, only have the signal handling capacity of the reliable registering capacity and part of critical data,
Complicated fault diagnosis and health forecast method, which is then placed in the PHM systems of ground, to be executed so that the fault diagnosis of equipment and repair
It is recommended that can not provide in real time or in time, cause that equipment safety operation guarantee is delayed or is delayed.
Invention content
In order to solve the above technical problem, the present invention provides a kind of prediction health control methods and device, computer can
Storage medium is read, the arithmetic speed of equipment can be improved.
In order to reach the object of the invention, what the technical solution of the embodiment of the present invention was realized in：
An embodiment of the present invention provides a kind of prediction health control methods, including：
The data of sensor acquisition are obtained, chronologically adding window is divided into multiple data blocks by the data of acquisition；
Concurrent Feature extraction, and the feature that each thread is extracted are carried out to the multiple data blocks marked off by multithreading
Category merges, the characteristic data set after being merged；
Health evaluating, fault diagnosis or predicting residual useful life are carried out using obtained characteristic data set.
Further, when executing the multithreading by multiple multiprocessor MP, between the multithreading inside identical MP
By shared drive into row data communication；By global memory into row data communication between different MP.
Further, described to carry out health evaluating using obtained characteristic data set, including：
The characteristic data set is divided into multiple Sub Data Sets；
Cluster operation is carried out at the same time to each Sub Data Set of division by multithreading, obtains the part of each Sub Data Set
Cluster result；
The Local Clustering result of each Sub Data Set is subjected to merger processing, obtains the cluster result after merger.
Further, described to carry out fault diagnosis or predicting residual useful life using obtained characteristic data set, including：
The characteristic data set is inputted to the depth confidence neural network model pre-established；
By multi-threaded parallel Gibbs model, mutually it is activated according between sampling results parallel computation hidden layer and aobvious layer
Probability, update the weight and bias between each neuron parallel；
Using between updated each neuron weight and bias as initialization training parameter, to depth confidence
Neural network is with having carried out supervision trained；
Fault diagnosis or predicting residual useful life are carried out using trained depth confidence neural network.
Further, described by multi-threaded parallel Gibbs model, according to sampling results parallel computation hidden layer and aobvious layer
Between the probability that is mutually activated, update the weight and bias between each neuron parallel, including：
Parallel Gibbs model, while the value for showing each neuron on layer is extracted, each hidden neuron is calculated by aobvious layer
The probability value of neuronal activation：
Wherein i ∈ [0,M],j∈[0,N], v is aobvious layer data value, and h is hidden layer data value, and M is aobvious layer neuron number, N
For hidden neuron number；W is weight, and b is aobvious layer bias, and c is hidden layer bias；
Pass through hidden layer h(0)Reconstruct shows layer v(1), parallel Gibbs model, calculates and shows layer neuron by hidden neuron again
The probability value of activation：
By showing layer v(1)Reconstruct hidden layer h(1), parallel Gibbs model, calculates hidden neuron by aobvious layer neuron again
The probability value of activation：
The weight and bias between each neuron are updated parallel：
Wherein, η is preset learning rate.
Further, the weight and bias using between updated each neuron is as initialization training ginseng
Number, it is with having carried out supervision trained to depth confidence neural network, including：
According to back-propagation algorithm, supervision ground has been carried out to the depth B P neural networks in the depth confidence neural network
Tuning is trained；
The weight and bias that each layer is updated using gradient descent algorithm iteration, until the weight and bias of each layer reach
It optimizes.
The embodiment of the present invention additionally provides a kind of computer readable storage medium, is deposited on the computer readable storage medium
One or more program is contained, one or more of programs can be executed by one or more processor, to realize such as
The step of prediction health control method described in any of the above item.
The embodiment of the present invention additionally provides a kind of prediction health controller, including feature extraction unit and prediction health pipe
Unit is managed, wherein：
Feature extraction unit, for obtain sensor acquisition data, by the data of acquisition chronologically adding window be divided into it is more
A data block；Concurrent Feature extraction, and the spy that each thread is extracted are carried out to the multiple data blocks marked off by multithreading
Sign category merges, the characteristic data set after being merged；
Predict health control unit, the characteristic data set for being obtained using feature extraction unit carries out health evaluating, event
Barrier diagnosis or predicting residual useful life.
Further, described to predict being good for using the characteristic data set that feature extraction unit obtains for health control unit
Health is assessed, including：
The characteristic data set that the feature extraction unit obtains is divided into multiple Sub Data Sets；By multithreading to dividing
Each Sub Data Set be carried out at the same time cluster operation, obtain the Local Clustering result of each Sub Data Set；By each Sub Data Set
Local Clustering result carry out merger processing, obtain the cluster result after merger.
Further, the characteristic data set of the prediction health control unit obtained using feature extraction unit carries out event
Barrier diagnosis or predicting residual useful life, including：
The characteristic data set that the feature extraction unit obtains is inputted into the depth confidence neural network model pre-established；
By multi-threaded parallel Gibbs model, according to the probability being mutually activated between sampling results parallel computation hidden layer and aobvious layer,
The weight and bias between each neuron are updated parallel；By between updated each neuron weight and biasing
Value is with having carried out supervision trained to depth confidence neural network as initialization training parameter；Utilize trained depth confidence
Neural network carries out fault diagnosis or predicting residual useful life.
Technical scheme of the present invention has the advantages that：
Prediction health control method and device, computer readable storage medium provided by the invention, by adopting sensor
The data time sequence windowing process of collection is divided into multiple data blocks, then carries out Concurrent Feature extraction, provides a kind of support parallel
The health control platform of operation improves the upper PHM system operations speed of equipment and information analysis ability, to realize real-time failure
Diagnosis and repair are suggested generating, and have effectively ensured the safe operation of equipment.
Description of the drawings
Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, this hair
Bright illustrative embodiments and their description are not constituted improper limitations of the present invention for explaining the present invention.In the accompanying drawings：
Fig. 1 is a kind of flow diagram of prediction health control method of the embodiment of the present invention；
Fig. 2 is a kind of open parallel computation isomery framework PHM system hierarchy schematic diagrames of the embodiment of the present invention；
Fig. 3 is the feature extraction concurrent operation schematic diagram of the embodiment of the present invention；
Fig. 4 is the feature extraction parallel computation GPU method schematic diagrams of the embodiment of the present invention；
Fig. 5 is that the K-means of the embodiment of the present invention clusters concurrent operation flow diagram；
Fig. 6 is that the K-means of the embodiment of the present invention clusters concurrent operation GPU flow diagrams；
Fig. 7 is in the related technology by notch hereby graceful machine simple structure schematic diagram；
Fig. 8 is depth confidence network (Deep Belief Nets, DBN) concurrent operation flow chart of the embodiment of the present invention；
Fig. 9 is the DBN principle schematic diagrams of the embodiment of the present invention；
Figure 10 is a kind of structural schematic diagram of prediction health controller of the embodiment of the present invention.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention
Embodiment be described in detail.It should be noted that in the absence of conflict, in the embodiment and embodiment in the application
Feature mutually can arbitrarily combine.
As shown in Figure 1, a kind of prediction health control method according to the present invention, includes the following steps：
Step 101：The data of sensor acquisition are obtained, chronologically adding window is divided into multiple data blocks by the data of acquisition；
It should be noted that adding window of the present invention refers to carrying out piecemeal to data in time domain.Sensor collects
Data chronologically handle the data block that adding window is divided into multiple equal sizes, data block is stored in being total to for processor in each window
It enjoys in memory, is used for processor multithreading is shared.
Step 102：Concurrent Feature extraction is carried out to the multiple data blocks marked off by multithreading, and each thread is carried
The feature category taken merges, the characteristic data set after being merged；
It should be noted that the hardware of the upper PHM systems of equipment can be by central processing unit (Central Processing
Unit, CPU)+graphics processor (Graphics Processing Unit, GPU) or CPU+ field programmable gate arrays
(Field-Programmable Gate Array, FPGA) two kinds of system platform structure compositions, CPU run primary control program, GPU
As data parallel equipment, operation needs the algorithm routine of concurrent operation, using general purpose GPU (General
Purpose GPU, GPGPU) technology and open operation language (Open Computing Language, OpenCL) technology complete
General-purpose computations.FPGA system is realized to be solved by the concurrent operation of algorithm, c program (hardware c program) to FPGA target data files
Scheme.PHM programs realize diagnosis, prediction and health state evaluation, the parallel computation for passing through standard using the algorithm of parallel computation
Framework drives bottom hardware so that diagnosis prediction program is realized independently of bottom hardware, can be common to using GPU's or FPGA
Embedded computing system.
Fig. 2 is a kind of the health control plateform system hierarchical structure of concurrent operation to be supported to illustrate according to what the present invention realized
Figure, the bottom is hardware configuration, is made of CPU, GPU or FPGA, the primary control program that CPU operation orders execute, GPU (or FPGA)
The algorithm routine of application acceleration technology concurrent operation is run, completes to accelerate general purpose operation.Middle layer is operating system and simultaneously
Row operational software framework, operating system complete the functions such as program scheduler, memory management, human-computer interaction, parallel on CPU
Operational software framework mainly applies GPGPU and OpenCL technologies to realize from health control algorithm routine to GPU concurrent operation programs
Realization.Top layer is the algorithm routine applied to health control, include mainly data processing features extraction (root mean square operation,
The flexure factor, kurtosis factor etc.), health evaluating (K mean values (K-means) cluster, hidden Markov model etc.), fault diagnosis
(depth confidence neural network, support vector machines etc.), predicting residual useful life (hidden Markov model, neural network etc.).
The present invention mainly applies the numerous processing cores of GPU and higher bandwidth of memory and stronger processing highly-parallel
Data-intensive computing ability and open, compatible, the free standard towards the general purpose multiple programming of heterogeneous system
OpenCL opens operation language, and concurrent operation is realized in CPU, GPU, FPGA heterogeneous system.By using open parallel computation frame
Structure, the present invention provides a kind of health control platforms for supporting concurrent operation, improve the upper PHM system operations speed of equipment and letter
Breath analysis ability has effectively ensured the safe operation of equipment to realize that real-time fault diagnosis and repair suggest generating.
In PHM systems, sensor (measuring each system component status signal) data on equipment are transferred to PHM systems,
Data processing and information analysis main operational, which are feature extraction processing and health control analysis, the present invention, realizes these cores
The concurrent operation platform of operation.In our CPU+GPU heterogeneous systems, CPU is responsible for the logic of the whole process of serial process
Control and serial task therein are organized organization data and are copied in the memory of GPU, then GPU called to generate multiple threads
And be mapped on the computing array of GPU, kernel (Kernel) function is executed parallel with OpenCL technical controlling GPU many-cores on GPU,
Concurrent operation is carried out, GPU is finished, and data are copied go back to the ends CPU by serial program, complete once-through operation.The operation list on GPU
Data communication between member passes through global memory by Sharing Memory Realization in GPU blocks between GPU blocks.Hereinafter, being directed to feature respectively
For extraction and health evaluating clustering method, illustrate concurrent operation software desk Implementation method.
Further, identical when executing the multithreading by multiple multiprocessors (Multi Processor, MP)
By shared drive into row data communication between multithreading inside MP；Data are carried out between different MP by global memory to lead to
Letter.
In multiprocessor, start multithreading simultaneously with each MP on OpenCL technical controllings GPU, it is parallel to execute kernel in MP
Function, each kernel functions (time domain, time and frequency zone signal characteristic abstraction function) call the data in same shared drive, warp
The data crossed after data processing are deposited into the global store of GPU.Inventive novelty herein is due to PHM feature extractions
Sequential processing, shared data is with uniformity to the data of multithreading operation in MP, and at the same time carrying out the multichannel feature of signal
Operation.Wherein kernel functions are signal characteristic abstraction function, they are the root mean square of time domain, the flexure factor, the kurtosis factor, wave
The peak factor, the nargin factor, shape factor, pulse factor operation function and time-frequency domain empirical mode decomposition eigen mode letter
Number.After feature extraction and calculation, uniformity signal characteristic is grouped together in global memory.
Signal characteristic abstraction concurrent operation method is as shown in figure 3, multiple characteristic operations execute parallel.
Its GPU implementation method is as shown in Figure 4, wherein Rms is root mean square operation, and Cf is crest factor operation, and If is pulse
Factor operation, Mf are nargin factor operations, and Sk is flexure factor operation, and Ku is kurtosis factor operation, and Sf is shape factor operation,
Emd is empirical mode decomposition operation.
Step 103：Health evaluating, fault diagnosis or predicting residual useful life are carried out using obtained characteristic data set.
It should be noted that the collected initial data of sensor generates characteristic data set after signal characteristic abstraction,
Health state evaluation and fault diagnosis for PHM systems.Prediction health control method in the present embodiment is using parallel special
The method of sign extraction extracts characteristic data set.Step 103 can carry out prediction health control using the method for serial or parallel.
Further, described to carry out health evaluating using obtained characteristic data set, including：
The characteristic data set is divided into multiple Sub Data Sets；
Cluster operation is carried out at the same time to each Sub Data Set of division by multithreading, obtains the part of each Sub Data Set
Cluster result；
The Local Clustering result of each Sub Data Set is subjected to merger processing, obtains the cluster result after merger.
Further, the cluster operation is that K-means clusters operation.
For health state evaluation, the non-supervisory automatic cluster to characteristic data set is carried out using K-means clusterings,
To generate the data acquisition system (being often divided into five classes, health, can be used, failure, scrap inferior health) of different Health Categories.
Wherein K-means algorithm principles are as follows：
Assuming that data set D={ x1,x2,…xm, cluster numbers K
1) k sample is randomly choosed from data set D as initial mean value vector { u1,u2,…..,uk}。
2) each sample x is calculatedjWith each mean vector uiThe distance of (1≤i≤k), formula are：
dji=||xj-ui||2
According to sample xjSample x is determined with the minimum distance size of mean vectorjCluster a small bundle of straw, etc. for silkworms to spin cocoons on label, by sample xjIt draws
Enter to corresponding cluster.
3) mean vector of each cluster a small bundle of straw, etc. for silkworms to spin cocoons on is calculated, and updates mean vector value.
4) cycle executes the 2) step and the 3) step, until mean vector stops updating.
5) finally, data set is divided into K classes.
Signal characteristic data set as shown in figure 5, is divided into the subnumber of multiple same sizes by K-means concurrent operations flow
According to collection, K-means cluster operations are carried out at the same time, the data after clustering carry out aggregation of data processing, finally export merger again
Data afterwards.
Characteristic set as shown in fig. 6, is divided into the smaller characteristic of multiple equal sizes by GPU concurrent operations method
According to collection, these smaller characteristic data sets are stored in the shared drive in each MP, each MP starts line simultaneously in GPU blocks
Journey is performed simultaneously K-means in MP and clusters operation to be carried out at the same time Local Clustering to the conjunction of these small data sets.
Specifically, the health state evaluation level results data of clustering obtained according to cluster operation are in the overall situation
Middle carry out aggregation of data is deposited, i.e., the identical corresponding characteristic data set of health state evaluation grade is grouped together, and is finally exported
The result of merger data.
Further, described to carry out fault diagnosis or predicting residual useful life using obtained characteristic data set, including：
The characteristic data set is inputted to the depth confidence neural network model pre-established；
By multi-threaded parallel Gibbs model, mutually it is activated according between sampling results parallel computation hidden layer and aobvious layer
Probability, update the weight and bias between each neuron parallel；
Using between updated each neuron weight and bias as initialization training parameter, to depth confidence
Neural network is with having carried out supervision trained；
Fault diagnosis or predicting residual useful life are carried out using trained depth confidence neural network.
For the fault diagnosis of PHM systems, carried out using DBN (the depth confidence neural network based on deep neural network)
Automatic classification to characteristic data set, to carry out fault diagnosis.Deep neural network using RBM (by notch hereby graceful machine), and
Row operation design is as follows：
RBM by notch hereby graceful machine simple structure schematic diagram as shown in fig. 7, wherein each limited Boltzmann machine is a kind of god
Through perceptron, it is made of with hidden layer an aobvious layer.It is two-way between hidden layer H and aobvious layer V neurons to connect.
RBM algorithm principles：
The probability that hidden neuron is shown layer neuronal activation is：P(Hj|V)=sigmoid (bj+∑iwi,j*xi)
Show layer neuron is by the probability that hidden neuron activates：
P(V|Hj)=sigmoid (cj+∑jwi,j*Hi)
Wherein w is bonding strength, and b is the biasing coefficient to showing layer, and c is the biasing coefficient to hidden layer.
xiTo show layer neuron wherein i=1,2,3..NV, HjFor hidden neuron wherein j=1,2,3 ... NH。
The probability that each hidden neuron is activated is calculated, and threshold value u is set, and the value range of u is 0-1, if more than
Threshold value is then activated, and is not otherwise activated.The activation probability for similarly calculating each aobvious layer neuron, judges whether it is activated.
First, data are passed into aobvious layer, calculates the probability P (H that each neuron of hidden layer is activated1|V1).From the general of calculating
Gibbs model is taken to extract a sample, H in rate distribution1~P (H1|V1)
Use H1It rebuilds and shows layer, aobvious layer is pushed away by the way that hidden layer is counter.
In next step, the probability P (V for showing that each neuron is activated in layer is calculated2|H1), similarly taken out using Gibbs model
Take a sample, V2~P (V2|H1)
In next step, the probability that each hidden neuron is activated is calculated, distribution P (H are obtained2|V2)
In next step, weight is updated.W=w+ η * (P (H1|V1)*V1-P(H2|V2)*V2)
B=b+ η * (V1-V2)
C=c+ η * (H1-H2)
Wherein η is learning rate.
Collateral learning gathers better than the part that existing learning process is multichannel Non-surveillance clustering while handling small data set
Class, to accelerate the speed of overall data clustering.
As shown in figure 8, in RBM concurrent operations design, include the following steps：
First, each neuron of parallel computation hidden layer is activated probability,
In next step, it according to Gibbs model, is connected as sample calculating in hidden layer paralleling abstracting neuron and is each shown
Layer neuron is activated probability, pushes away aobvious layer neuron using hidden neuron is counter and is activated probability；
It should be noted that parallel Gibbs model herein is each neuron while being extracted, and calculate
Aobvious layer neuron is activated probability, and Gibbs model in the prior art is sequentially to extract neuron one by one and calculate its aobvious layer
Neuron is activated probability.
In next step, layer neuron is shown using parallel Gibbs model, calculates the aobvious anti-hidden neuron that pushes away of layer neuron and is swashed
Probability living
Finally, weight is updated parallel.
Parallel update weight herein, newer is the connection weight and bias between hidden neuron and aobvious layer
(and being to update simultaneously), and update weight in the prior art is to update the company of hidden neuron and aobvious layer neuron one by one
Connect weight and bias.
DBN principle schematic diagrams are as shown in Figure 9：
Depth confidence neural network is made of multiple limited Boltzmann machines (RBM) with depth B P neural networks.Depth
The training of confidence neural network model is mainly following two steps：
1) each layer of unsupervised training is limited Boltzmann's network (RBM)
Training characteristics data input, and initialize weighted value w between the hidden layer of each RBM and aobvious layer and each other neuron
Between bias b, c.It calculates between hidden layer and aobvious layer, be mutually activated probability.
RBM networks in DBN neural networks can make the initialization weighting parameter of deep layer BP neural network, make entire DBN
Neural network compensates for because training time for being brought because random parameter initializes of BP neural network is long, and is easily absorbed in office
The shortcomings of portion is optimal.
2) training parameter obtained after the unsupervised training of RBM trains ginseng as the initialization of depth B P neural networks
Number, it is with having carried out supervision trained
In order to reach the desired value of trained label, according to back-propagation algorithm, carry out for the depth in entire neural network
Spend training weights in BP neural network, the small parameter perturbations of bias, and the training weighted value of each layer for being optimal, biasing
Value.
Parallel DBN algorithm principles：
First, characteristic is input to aobvious layer.Random initializtion weight w is b to aobvious layer bias, to hidden layer bias c
Parameter.
It calculates aobvious layer data value v and hidden layer data value h and passes through formula：
H=w*v+c；
V=wT*h+b；
Wherein, v is the matrix of M*1, and h is the matrix of N*1, and M is aobvious layer neuron number, and N is hidden neuron number.W
For the matrix of M*N, b is the matrix of M*1, and c is the matrix of N*1.
In next step, parallel Gibbs model extracts each neuron shown on layer simultaneously, calculated with its corresponding value
Each hidden neuron is shown the probability value of layer neuronal activation, and the probability that hidden neuron is shown layer neuronal activation is：
Wherein i ∈ [0,M],j∈[0,N]
In next step, pass through hidden layer h(0)Reconstruct shows layer v(1), parallel Gibbs model, calculates and shows what layer was activated by hidden layer again
Probability value, aobvious layer neuron are by the probability that hidden neuron activates：
In next step, by showing layer v(1)Reconstruct hidden layer h(1), parallel Gibbs model, calculates hidden layer and is shown layer activation again
Probability value, the probability that hidden neuron is shown layer neuronal activation are：
In next step, the weight between update neuron and biasing parallel
Wherein, η is learning rate.
In next step, hidden layer output is calculated according to trained w, b and uses formula：hl=sigmoid (wl*v(2)+bl)；
H herein represents the output of hidden layer, and l is the index number of the hidden layer of BP neural network, v(2)As BP nerve nets
The input layer of network.
In next step, the output valve y of output layer is calculated：Y=sigmoid (h), wherein y is the matrix of M*1.
In next step, the model parameter of network is updated using the back-propagation algorithm of least-mean-square-error criterion, has carried out prison
The tuning training superintended and directed.Input expected matrix (training label matrix) y '.Calculate cost function：
Wherein, y ' is the matrix of M*1, and E is the matrix of M*1.
In next step, using gradient descent algorithm, iteration updates weight and the biasing of each layer：
Until E → 0, stop update weight and bias.
According to trained each layer weighted value and bias, calculates and export final classification results.
The embodiment of the present invention additionally provides a kind of computer readable storage medium, is deposited on the computer readable storage medium
One or more program is contained, one or more of programs can be executed by one or more processor, to realize such as
The step of prediction health control method described in any of the above item.
As shown in Figure 10, an embodiment of the present invention provides a kind of prediction health controllers, including feature extraction unit
1001 and prediction health control unit 1002, wherein：
Feature extraction unit 1001, the data for obtaining sensor acquisition, by the data of acquisition, chronologically adding window divides
For multiple data blocks；Concurrent Feature extraction is carried out to the multiple data blocks marked off by multithreading, and each thread is extracted
Feature category merge, the characteristic data set after being merged；
Predict health control unit 1002, the characteristic data set for being obtained using feature extraction unit 1001 carries out health
Assessment, fault diagnosis or predicting residual useful life.
Further, when feature extraction unit 1001 executes the multithreading by multiple multiprocessor MP, identical MP
By shared drive into row data communication between internal multithreading；By global memory into row data communication between different MP.
Specifically, the collected data of sensor chronologically handle the data block that adding window is divided into multiple equal sizes, often
Data block is stored in the shared drive of processor in a window, is used for processor multithreading is shared；Passing through feature extraction meter
After calculation, uniformity signal characteristic is grouped together in global memory.
Further, feature extraction unit 1001 extract feature include the root mean square of time domain, the flexure factor, kurtosis because
Son, crest factor, the nargin factor, shape factor, pulse factor operation function and time-frequency domain empirical mode decomposition it is intrinsic
Modular function.
Further, the feature for being obtained using feature extraction unit 1001 of the prediction health control unit 1002
Data set carries out health evaluating, including：
The characteristic data set that feature extraction unit 1001 obtains is divided into multiple Sub Data Sets；By multithreading to dividing
Each Sub Data Set be carried out at the same time cluster operation, obtain the Local Clustering result of each Sub Data Set；By each Sub Data Set
Local Clustering result carry out merger processing, obtain the cluster result after merger.
Further, the cluster operation is that K-means clusters operation.
Further, when predicting that health control unit 1002 executes the multithreading by multiple multiprocessor MP, phase
With between the multithreading inside MP by shared drive into row data communication；Data are carried out between different MP by global memory to lead to
Letter.
Specifically, characteristic set is divided into the smaller feature of multiple equal sizes by prediction health control unit 1002
These smaller characteristic data sets are stored in the shared drive in each MP by data set, and each MP starts line simultaneously in GPU blocks
Journey is performed simultaneously K-means in MP and clusters operation to be carried out at the same time Local Clustering to the conjunction of these small data sets；By clustering
Data aggregation of data is carried out in global memory according to the obtained health state evaluation level results of cluster operation, i.e., it is identical strong
The corresponding characteristic data set of health status assessment grade is grouped together, and finally exports the result of merger data.
Further, the feature for being obtained using feature extraction unit 1001 of the prediction health control unit 1002
Data set carries out fault diagnosis or predicting residual useful life, including：
The characteristic data set that feature extraction unit 1001 obtains is inputted into the depth confidence neural network model pre-established；
By multi-threaded parallel Gibbs model, according to the probability being mutually activated between sampling results parallel computation hidden layer and aobvious layer,
The weight and bias between each neuron are updated parallel；By between updated each neuron weight and biasing
Value is with having carried out supervision trained to depth confidence neural network as initialization training parameter；Utilize trained depth confidence
Neural network carries out fault diagnosis or predicting residual useful life.
Further, the prediction health control unit 1002 by multi-threaded parallel Gibbs model, according to sampling
As a result the probability being mutually activated between parallel computation hidden layer and aobvious layer updates weight between each neuron and partially parallel
Value is set, including：
Parallel Gibbs model, while the value for showing each neuron on layer is extracted, each hidden neuron is calculated by aobvious layer
The probability value of neuronal activation：
Wherein i ∈ [0,M],j∈[0,N], v is aobvious layer data value, and h is hidden layer data value, and M is aobvious layer neuron number, N
For hidden neuron number；W is weight, and b is aobvious layer bias, and c is hidden layer bias；
Pass through hidden layer h(0)Reconstruct shows layer v(1), parallel Gibbs model, calculates and shows layer neuron by hidden neuron again
The probability value of activation：
By showing layer v(1)Reconstruct hidden layer h(1), parallel Gibbs model, calculates hidden neuron by aobvious layer neuron again
The probability value of activation：
The weight and bias between each neuron are updated parallel：
Wherein, η is preset learning rate.
Further, the prediction health control unit 1002 by between updated each neuron weight and
Bias is used as initialization training parameter, with having carried out supervision trained to depth confidence neural network, including：
According to back-propagation algorithm, supervision ground has been carried out to the depth B P neural networks in the depth confidence neural network
Tuning is trained；
The weight and bias that each layer is updated using gradient descent algorithm iteration, until the weight and bias of each layer reach
It optimizes.
The present invention realizes the general PHM parallel computing architectures framework of cross-platform isomery, is realized independently of bottom hardware, can
To support CPU, GPU and FPGA various processor platform；Using general purpose GPU and open parallel computation architecture technology, have
Improve PHM system operations ability and real-time to effect so that more complicated health control algorithm can be handled on equipment, and
The fault diagnosis and repair for providing equipment in time are suggested.
One of ordinary skill in the art will appreciate that all or part of step in the above method can be instructed by program
Related hardware is completed, and described program can be stored in computer readable storage medium, such as read-only memory, disk or CD
Deng.Optionally, all or part of step of above-described embodiment can also be realized using one or more integrated circuits, accordingly
Ground, the form that hardware may be used in each module/unit in above-described embodiment are realized, the shape of software function module can also be used
Formula is realized.The present invention is not limited to the combinations of the hardware and software of any particular form.
Although describing the invention in detail above, the foregoing is merely the preferred embodiment of the present invention
, it is not intended to restrict the invention, for those skilled in the art, the invention may be variously modified and varied.It is all
Within the spirit and principles in the present invention, any modification, equivalent replacement, improvement and so on should be included in the guarantor of the present invention
Within the scope of shield.
Claims (10)
1. a kind of prediction health control method, which is characterized in that including：
The data of sensor acquisition are obtained, chronologically adding window is divided into multiple data blocks by the data of acquisition；
Concurrent Feature extraction is carried out to the multiple data blocks marked off by multithreading, and the feature that each thread is extracted is by class
It does not merge, the characteristic data set after being merged；
Health evaluating, fault diagnosis or predicting residual useful life are carried out using obtained characteristic data set.
2. prediction health control method according to claim 1, which is characterized in that executed when by multiple multiprocessor MP
When the multithreading, by shared drive into row data communication between the multithreading inside identical MP；By complete between different MP
Intra-office deposits into row data communication.
3. prediction health control method according to claim 1, which is characterized in that described to utilize obtained characteristic data set
Health evaluating is carried out, including：
The characteristic data set is divided into multiple Sub Data Sets；
Cluster operation is carried out at the same time to each Sub Data Set of division by multithreading, obtains the Local Clustering of each Sub Data Set
As a result；
The Local Clustering result of each Sub Data Set is subjected to merger processing, obtains the cluster result after merger.
4. prediction health control method according to claim 1, which is characterized in that described to utilize obtained characteristic data set
Fault diagnosis or predicting residual useful life are carried out, including：
The characteristic data set is inputted to the depth confidence neural network model pre-established；
It is general according to what is be mutually activated between sampling results parallel computation hidden layer and aobvious layer by multi-threaded parallel Gibbs model
Rate updates weight and bias between each neuron parallel；
Using between updated each neuron weight and bias as initialization training parameter, to depth confidence nerve
Network is with having carried out supervision trained；
Fault diagnosis or predicting residual useful life are carried out using trained depth confidence neural network.
5. prediction health control method according to claim 4, which is characterized in that described to pass through multi-threaded parallel gibbs
Sampling, according to the probability being mutually activated between sampling results parallel computation hidden layer and aobvious layer, update parallel each neuron it
Between weight and bias, including：
Parallel Gibbs model, while the value for showing each neuron on layer is extracted, each hidden neuron is calculated by aobvious layer nerve
The probability value of member activation：
Wherein i ∈ [0, M], j ∈ [0, N], v is aobvious layer data value, and h is hidden layer data value, and M is aobvious layer neuron number, and N is hidden
Layer neuron number；W is weight, and b is aobvious layer bias, and c is hidden layer bias；
Pass through hidden layer h(0)Reconstruct shows layer v(1), parallel Gibbs model, calculates aobvious layer neuron and is activated by hidden neuron again
Probability value：
By showing layer v(1)Reconstruct hidden layer h(1), parallel Gibbs model, calculates hidden neuron by aobvious layer neuronal activation again
Probability value：
The weight and bias between each neuron are updated parallel：
Wherein, η is preset learning rate.
6. prediction health control method according to claim 4, which is characterized in that described by updated each neuron
Between weight and bias as initialization training parameter, to depth confidence neural network carried out supervision it is trained, wrap
It includes：
According to back-propagation algorithm, supervision ground tuning has been carried out to the depth B P neural networks in the depth confidence neural network
Training；
The weight and bias that each layer is updated using gradient descent algorithm iteration, until the weight and bias of each layer are optimal
Change.
7. a kind of computer readable storage medium, which is characterized in that on the computer readable storage medium storage there are one or
The multiple programs of person, one or more of programs can be executed by one or more processor, with realize as claim 1 to
The step of prediction health control method described in any one of 6.
8. a kind of prediction health controller, which is characterized in that including feature extraction unit and predict health control unit,
In：
Feature extraction unit, the data for obtaining sensor acquisition, by the data of acquisition, chronologically adding window is divided into multiple numbers
According to block；Concurrent Feature extraction is carried out to the multiple data blocks marked off by multithreading, and the feature that each thread extracts is pressed
Classification merges, the characteristic data set after being merged；
Predict health control unit, the characteristic data set for being obtained using feature extraction unit carries out health evaluating, failure is examined
Disconnected or predicting residual useful life.
9. prediction health controller according to claim 8, which is characterized in that the profit of the prediction health control unit
Health evaluating is carried out with the characteristic data set that feature extraction unit obtains, including：
The characteristic data set that the feature extraction unit obtains is divided into multiple Sub Data Sets；By multithreading to each of division
A Sub Data Set is carried out at the same time cluster operation, obtains the Local Clustering result of each Sub Data Set；By the office of each Sub Data Set
Portion's cluster result carries out merger processing, obtains the cluster result after merger.
10. prediction health controller according to claim 8, which is characterized in that the prediction health control unit
Fault diagnosis or predicting residual useful life are carried out using the characteristic data set that feature extraction unit obtains, including：
The characteristic data set that the feature extraction unit obtains is inputted into the depth confidence neural network model pre-established；Pass through
Multi-threaded parallel Gibbs model, according to the probability being mutually activated between sampling results parallel computation hidden layer and aobvious layer, parallel
Update the weight and bias between each neuron；By the weight and bias work between updated each neuron
It is with having carried out supervision trained to depth confidence neural network to initialize training parameter；Utilize trained depth confidence nerve
Network carries out fault diagnosis or predicting residual useful life.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810469967.2A CN108710958B (en) | 2018-05-16 | 2018-05-16 | Predictive health management method and device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810469967.2A CN108710958B (en) | 2018-05-16 | 2018-05-16 | Predictive health management method and device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108710958A true CN108710958A (en) | 2018-10-26 |
| CN108710958B CN108710958B (en) | 2022-04-15 |
Family
ID=63868182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810469967.2A Active CN108710958B (en) | 2018-05-16 | 2018-05-16 | Predictive health management method and device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108710958B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109445970A (en) * | 2018-09-18 | 2019-03-08 | 北京工业大学 | A kind of software reliability Time Series Forecasting Methods and application |
| CN111795824A (en) * | 2019-04-09 | 2020-10-20 | 中国航发商用航空发动机有限责任公司 | Fault diagnosis method and system for mechanical power system |
| CN116579762A (en) * | 2023-04-14 | 2023-08-11 | 广州林旺空调工程有限公司 | Intelligent operation and maintenance platform for cooling tower |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104849698A (en) * | 2015-05-21 | 2015-08-19 | 中国人民解放军海军工程大学 | Radar signal parallel processing method and system based on heterogeneous multinucleated system |
| US20160071010A1 (en) * | 2014-05-31 | 2016-03-10 | Huawei Technologies Co., Ltd. | Data Category Identification Method and Apparatus Based on Deep Neural Network |
| CN106405640A (en) * | 2016-08-26 | 2017-02-15 | 中国矿业大学(北京) | Automatic microseismic signal arrival time picking method based on depth belief neural network |
| US20170102984A1 (en) * | 2015-10-13 | 2017-04-13 | Huawei Technologies Co., Ltd. | Fault Diagnosis Method and Apparatus for Big-Data Network System |
| CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Self adaptation depth confidence network Method for Bearing Fault Diagnosis based on Nesterov momentum methods |
| CN107101828A (en) * | 2017-05-27 | 2017-08-29 | 辽宁大学 | The bearing residual life Forecasting Methodology that a kind of encoder health degree is evaluated |
| CN107274011A (en) * | 2017-06-05 | 2017-10-20 | 上海电力学院 | The equipment state recognition methods of comprehensive Markov model and probability net |
| CN107316078A (en) * | 2016-04-27 | 2017-11-03 | 北京中科寒武纪科技有限公司 | Apparatus and method for performing artificial neural network self study computing |
-
2018
- 2018-05-16 CN CN201810469967.2A patent/CN108710958B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160071010A1 (en) * | 2014-05-31 | 2016-03-10 | Huawei Technologies Co., Ltd. | Data Category Identification Method and Apparatus Based on Deep Neural Network |
| CN104849698A (en) * | 2015-05-21 | 2015-08-19 | 中国人民解放军海军工程大学 | Radar signal parallel processing method and system based on heterogeneous multinucleated system |
| US20170102984A1 (en) * | 2015-10-13 | 2017-04-13 | Huawei Technologies Co., Ltd. | Fault Diagnosis Method and Apparatus for Big-Data Network System |
| CN107316078A (en) * | 2016-04-27 | 2017-11-03 | 北京中科寒武纪科技有限公司 | Apparatus and method for performing artificial neural network self study computing |
| CN106405640A (en) * | 2016-08-26 | 2017-02-15 | 中国矿业大学(北京) | Automatic microseismic signal arrival time picking method based on depth belief neural network |
| CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Self adaptation depth confidence network Method for Bearing Fault Diagnosis based on Nesterov momentum methods |
| CN107101828A (en) * | 2017-05-27 | 2017-08-29 | 辽宁大学 | The bearing residual life Forecasting Methodology that a kind of encoder health degree is evaluated |
| CN107274011A (en) * | 2017-06-05 | 2017-10-20 | 上海电力学院 | The equipment state recognition methods of comprehensive Markov model and probability net |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109445970A (en) * | 2018-09-18 | 2019-03-08 | 北京工业大学 | A kind of software reliability Time Series Forecasting Methods and application |
| CN111795824A (en) * | 2019-04-09 | 2020-10-20 | 中国航发商用航空发动机有限责任公司 | Fault diagnosis method and system for mechanical power system |
| CN116579762A (en) * | 2023-04-14 | 2023-08-11 | 广州林旺空调工程有限公司 | Intelligent operation and maintenance platform for cooling tower |
| CN116579762B (en) * | 2023-04-14 | 2023-10-20 | 广州林旺空调工程有限公司 | Intelligent operation and maintenance platform for cooling tower |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108710958B (en) | 2022-04-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108763360A (en) | A kind of sorting technique and device, computer readable storage medium | |
| Nabil et al. | Prediction of students’ academic performance based on courses’ grades using deep neural networks | |
| de Abril et al. | Connectivity inference from neural recording data: Challenges, mathematical bases and research directions | |
| CN106895975B (en) | Method for Bearing Fault Diagnosis based on Stacked SAE deep neural network | |
| Vanmali et al. | Using a neural network in the software testing process | |
| Chitra et al. | Review of heart disease prediction system using data mining and hybrid intelligent techniques | |
| CN109659033A (en) | A kind of chronic disease change of illness state event prediction device based on Recognition with Recurrent Neural Network | |
| Samhan et al. | Classification of Alzheimer's disease using convolutional neural networks | |
| Almadhoun et al. | Detection of brain tumor using deep learning | |
| CN110136842A (en) | Morbidity prediction technique, device and the computer readable storage medium of acute infectious disease | |
| Anand et al. | Fractional-Iterative BiLSTM Classifier: A Novel Approach to Predicting Student Attrition in Digital Academia | |
| CN108710958A (en) | A kind of prediction health control method and device, computer readable storage medium | |
| Das et al. | Detecting code smells using deep learning | |
| CN110363090A (en) | Intelligent heart disease detection method, device and computer readable storage medium | |
| CN110085314A (en) | Intelligent analysis method, system and the equipment of medical test data | |
| CN108647707A (en) | Probabilistic neural network creation method, method for diagnosing faults and device, storage medium | |
| CN113762486A (en) | Method and device for constructing fault diagnosis model of converter valve and computer equipment | |
| CN110310740A (en) | Based on see a doctor again information forecasting method and the system for intersecting attention neural network | |
| Es–SABERY et al. | An improved ID3 classification algorithm based on correlation function and weighted attribute | |
| Leonardi et al. | Explainable process trace classification: An application to stroke | |
| Saoud et al. | Application of data mining classification algorithms for breast cancer diagnosis | |
| Gao et al. | Research on software multiple fault localization method based on machine learning | |
| Rosnelly et al. | Tropical diseases identification using neural network adaptive resonance theory 2 | |
| Kadam et al. | A cnn-rnn based approach for simultaneous detection, identification and classification of intracranial hemorrhage | |
| Bhagwat et al. | Reviewing Machine Learning Algorithms in the Domain of Healthcare |
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 |