CN111382542B - Highway electromechanical device life prediction system facing full life cycle - Google Patents
Highway electromechanical device life prediction system facing full life cycle Download PDFInfo
- Publication number
- CN111382542B CN111382542B CN202010121280.7A CN202010121280A CN111382542B CN 111382542 B CN111382542 B CN 111382542B CN 202010121280 A CN202010121280 A CN 202010121280A CN 111382542 B CN111382542 B CN 111382542B
- Authority
- CN
- China
- Prior art keywords
- equipment
- electromechanical equipment
- highway
- electromechanical
- parameters
- 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.)
- Active
Links
- 238000006731 degradation reaction Methods 0.000 claims abstract description 51
- 238000004458 analytical method Methods 0.000 claims abstract description 43
- 230000015556 catabolic process Effects 0.000 claims abstract description 37
- 238000012216 screening Methods 0.000 claims abstract description 15
- 238000012544 monitoring process Methods 0.000 claims description 70
- 238000012423 maintenance Methods 0.000 claims description 61
- 238000000034 method Methods 0.000 claims description 49
- 230000036541 health Effects 0.000 claims description 32
- 230000006870 function Effects 0.000 claims description 30
- 230000007613 environmental effect Effects 0.000 claims description 29
- 239000013598 vector Substances 0.000 claims description 22
- 230000001186 cumulative effect Effects 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 13
- 208000033999 Device damage Diseases 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 9
- 230000015654 memory Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 125000004122 cyclic group Chemical group 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- MWUXSHHQAYIFBG-UHFFFAOYSA-N Nitric oxide Chemical compound O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 claims description 6
- 238000000513 principal component analysis Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 5
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 238000007689 inspection Methods 0.000 claims description 4
- 238000011160 research Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000007476 Maximum Likelihood Methods 0.000 claims description 2
- 238000005299 abrasion Methods 0.000 claims description 2
- 230000032683 aging Effects 0.000 claims description 2
- 230000005611 electricity Effects 0.000 claims description 2
- 230000004907 flux Effects 0.000 claims description 2
- 238000005286 illumination Methods 0.000 claims description 2
- 238000012417 linear regression Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 230000000737 periodic effect Effects 0.000 claims description 2
- 238000013138 pruning Methods 0.000 claims description 2
- 230000001360 synchronised effect Effects 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- 238000007726 management method Methods 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 8
- 238000010276 construction Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 102000008297 Nuclear Matrix-Associated Proteins Human genes 0.000 description 2
- 108010035916 Nuclear Matrix-Associated Proteins Proteins 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005315 distribution function Methods 0.000 description 2
- 230000003862 health status Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000008689 nuclear function Effects 0.000 description 2
- 210000000299 nuclear matrix Anatomy 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
Abstract
The invention provides a life prediction system of highway electromechanical equipment facing to a full life cycle, which comprises a highway tunnel key electromechanical equipment identification module, a road tunnel key electromechanical equipment identification module and a road tunnel prediction module, wherein the highway tunnel key electromechanical equipment identification module is used for identifying and screening key equipment in highway tunnel electromechanical equipment; the road tunnel electromechanical equipment residual life prediction module is used for predicting the residual life of key equipment identified and screened from the road tunnel electromechanical equipment; and the degradation analysis module is used for carrying out degradation analysis on the highway electromechanical equipment. The prediction system combines key equipment identification screening with the prediction of the residual life of the electromechanical equipment of the highway tunnel, and greatly improves the efficiency and accuracy of the prediction system. And by combining with degradation analysis of the highway electromechanical equipment, life prediction of the whole life cycle of the highway electromechanical equipment can be realized.
Description
Technical Field
The invention belongs to the field of road traffic electromechanical equipment, relates to life prediction of road electromechanical equipment, and in particular relates to a life prediction system of road electromechanical equipment for a full life cycle.
Background
The fault prediction and health management (Prognostics and Health Management, PHM) technology is a cross subject and a hot research direction covering multiple fields of basic materials, mechanical structures, energy sources, electronics, automatic tests, reliability, information and the like, can provide a comprehensive view of the health state of electromechanical equipment or the whole system through fault diagnosis, fault prediction, equipment residual service life (Remaining Useful Life, RUL) prediction and the like, and has been developed into an important support technology and foundation for system logistics guarantee, maintenance and autonomous health management in the aerospace field. Currently, fault diagnosis systems based on artificial intelligence are preliminarily replacing fault diagnosis systems based on conventional technologies due to their outstanding performance in improving the efficiency of monitoring technologies. Deep Learning (Deep Learning) is a branch of the most intelligent in the field of artificial intelligence, and is also a core topic and a popular research direction in the current machine Learning field. In face of various problems existing in the electromechanical system of the highway tunnel and equipment maintenance management, a PHM method based on deep learning is applied to provide a new thought and technical support for realizing intelligent management of the electromechanical system.
The existing life prediction framework of the road tunnel electromechanical equipment has the following problems: firstly, the importance degree of the tunnel electromechanical equipment is not subdivided, and is roughly divided according to the content of the highway tunnel maintenance standard, so that the differential maintenance management is not facilitated. Second, the influence of external factors on the degradation process of the equipment is not considered when the degradation model of the electromechanical equipment is established. Third, the existing life prediction model of the tunnel electromechanical device depends on characteristic data of the system, and is poor in adaptation to the situation that the road tunnel electromechanical device has intermittent faults.
Disclosure of Invention
Aiming at the defects and the shortcomings of the prior art, the invention aims to provide a life prediction system of highway electromechanical equipment for a full life cycle, which solves the technical problem that the life prediction result of the highway electromechanical equipment in the prior art is inaccurate.
In order to solve the technical problems, the invention adopts the following technical scheme:
the system comprises a highway tunnel electromechanical device residual life prediction module, wherein the highway tunnel electromechanical device residual life prediction module is used for predicting the residual life of the highway tunnel electromechanical device; the highway tunnel electromechanical device residual life prediction module comprises a highway tunnel electromechanical device residual life prediction model.
The invention also has the following technical characteristics:
the system also comprises a highway tunnel key electromechanical equipment identification module, wherein the highway tunnel key electromechanical equipment identification module is used for identifying and screening key equipment in highway tunnel electromechanical equipment; the residual life prediction module of the road tunnel electromechanical equipment is used for predicting the residual life of key equipment identified and screened from the road tunnel electromechanical equipment.
The system also comprises a degradation analysis module of the expressway electromechanical equipment, wherein the degradation analysis module of the expressway electromechanical equipment is used for carrying out degradation analysis on the expressway electromechanical equipment; the degradation analysis module of the highway electromechanical equipment comprises a degradation analysis model of the highway electromechanical equipment.
Compared with the prior art, the invention has the following technical effects:
and (I) the residual life prediction module of the road tunnel electromechanical equipment corrects and compensates the internal state parameters of the road tunnel electromechanical equipment by adopting external environment parameters, and improves the accuracy of input quantity in the residual life prediction model, thereby improving the accuracy of the trained model and the accuracy of the residual life obtained when the model is adopted for residual life prediction.
The invention takes the compensated internal state parameter and the failure rate of the road tunnel electromechanical equipment as analysis objects, obtains a feature vector through nuclear principal component analysis, takes the feature vector as the input quantity of a long-period memory network and can further improve the accuracy of the finally obtained residual life.
And (III) the residual life prediction module of the road tunnel electromechanical equipment establishes a tunnel electromechanical equipment operation environment state parameter acquisition module and a key performance parameter acquisition module, realizes the on-line monitoring and data acquisition of the road tunnel key electromechanical equipment, and obtains a complete original data set for the life prediction of the electromechanical equipment.
And (IV) constructing a fault model conforming to the failure rule of the electromechanical equipment of the highway tunnel by the residual life prediction module of the electromechanical equipment of the highway tunnel, qualitatively analyzing the life distribution of the electromechanical equipment, converting the running state time of the equipment into a fault rate as a main characteristic parameter, and constructing a life prediction characteristic vector of the electromechanical equipment of the highway tunnel by combining other relevant characteristic parameters.
And (V) effectively screening the characteristic parameters subjected to the environmental factor compensation treatment by adopting an unsupervised characteristic extraction method of nuclear main component analysis by using a residual life prediction module of the road tunnel electromechanical equipment.
The rest life prediction module of the road tunnel electromechanical equipment considers the characteristic that the life of the road tunnel electromechanical equipment has time-space correlation, and particularly provides a life prediction method of the road tunnel electromechanical equipment based on a cyclic neural network, which has better interpretability and higher accuracy.
And (VII) analyzing the degradation process of the highway electromechanical equipment by using the operation state monitoring data of the highway electromechanical equipment by using the degradation analysis module of the highway electromechanical equipment. Compared with the prior method for acquiring failure probability distribution of various equipment by analyzing the failure and maintenance data of the electromechanical equipment of the expressway, the method can grasp the degradation process of the equipment more accurately.
And (VIII) the degradation analysis module of the highway electromechanical equipment takes two factors of the running environment state and the maintenance state of the equipment closely related to the degradation process of the highway electromechanical equipment into consideration, and analyzes the influence mechanism of each factor on the degradation process of the equipment. By taking into account the effects of these factors on the degradation process of the device, the accuracy of the description of the degradation process of the device is improved.
(IX) the degradation analysis module of the electromechanical equipment of the expressway extracts the running state characteristics of the equipment through processing the state monitoring data of the electromechanical equipment of the expressway, and constructs a running state characteristic parameter set of the equipment. And then, screening the equipment degradation sensitive parameters according to the correlation degree of the parameters in the characteristic parameter set and the equipment health index parameters. And equipment degradation information in the equipment operation state monitoring data is effectively mined.
The (X) degradation analysis module of the electromechanical equipment of the expressway horizontally divides all the environmental factors by considering the influence mechanism of all the factors in the equipment operation environment on the degradation process of the equipment, and constructs parameters for representing the cumulative influence effect of all the environmental factors on the degradation process of the equipment. Meanwhile, historical data of hierarchical maintenance of the electromechanical equipment of the expressway is combined, and the accumulated influence effect parameters of the maintenance conditions are constructed. Compared with the traditional method for analyzing the equipment degradation process only from the equipment operation state monitoring data, the method further improves the accuracy of equipment degradation process description.
When the (XI) mechanical and electrical equipment degradation analysis module of the expressway is used for establishing an equipment degradation model by adopting a related vector machine algorithm, the nonlinear transformation processing of Box-Cox is carried out on the equipment health index parameters, and the accuracy of the model can be enhanced from the data processing angle.
The prediction system combines key equipment identification screening with the prediction of the residual life of the electromechanical equipment of the highway tunnel, and greatly improves the efficiency and accuracy of the prediction system. And by combining with degradation analysis of the highway electromechanical equipment, life prediction of the whole life cycle of the highway electromechanical equipment can be realized.
Drawings
Fig. 1 is an EQLC dataset sample.
Fig. 2 is a schematic diagram of a compensation flow of internal state parameters of the road tunnel electromechanical device based on gray correlation analysis according to the present invention.
Fig. 3 is a schematic diagram of a training process of the LSTM network of the present invention.
Fig. 4 is a schematic diagram of the LSTM network layer used in the present invention.
Fig. 5 is a measurement node in a PLC control box.
Fig. 6 is a sample graph of modeling results of a battery pack degradation process of a UPS power source device.
The following examples illustrate the invention in further detail.
Detailed Description
The following specific embodiments of the present invention are provided, and it should be noted that the present invention is not limited to the following specific embodiments, and all equivalent changes made on the basis of the technical solutions of the present application fall within the protection scope of the present invention.
Example 1:
the embodiment provides a highway tunnel key electromechanical equipment identification module, which comprises the following steps:
Step 1.1, acquiring the number gn of faults of the road tunnel electromechanical equipment in the monitored days, the equipment cost em and the maintenance cost wc according to the maintenance record of the road tunnel electromechanical equipment, and recording the maintenance grade drq of the road tunnel electromechanical equipment;
wherein:
if the road tunnel electromechanical equipment is replaced due to direct damage, the fault frequency gn of the road tunnel electromechanical equipment is still increased by 1;
when the maintenance process is daily inspection work on the appearance and the working state of the equipment, checking whether the working function, the performance and the running environment of the equipment meet the requirements or not and recording potential safety hazards, drq=1;
when the maintenance process is to carry out external periodic maintenance on the equipment; checking whether the shells of various electromechanical equipment leak electricity, and when the equipment is fastened, adjusted, derusted and preserved, drq=2;
when the maintenance process is to clean the inside of the equipment; maintaining the internal element, and fastening the internal connecting wire and the parts; when parts with serious abrasion are maintained and replaced and basic function tests of partial equipment are carried out, drq=3;
when the maintenance process is to perform overall maintenance and overhaul on equipment, including grounding test, core element inspection, fastening detection of wiring and parts, replacement of parts with serious damage or aging, and detection and debugging of main functions of the equipment, drq=4;
When the maintenance process is to comprehensively use technical means to test the key performance of the electromechanical equipment, the service life of the existing equipment, the use perfection rate of the equipment and the fault occurrence rate of the equipment are evaluated; drq=5 when evaluating the function and performance of the device;
step 1.2, evaluating the reliability of the computing device RE, re=gn/tm;
wherein tm is the number of days the device was monitored;
step 1.3, calculating the maintenance difficulty DE, maintenance cost and single price ratio BE of equipment; performing standardized treatment on the DE and the BE to obtain the result Dd and Bb, and calculating the maintenance evaluation WE of the equipment;
DE＝gn×dr q ；WE＝0.4Dd+0.6Bb；
step 1.4, calculating a loss cost QE of the device, qe=0.5ew+0.5mw, wherein:
the EW is an influence weight of the road tunnel electromechanical device damage to the system function, and when the influence of the road tunnel electromechanical device damage to the system function is no influence, the ew=1, when the influence of the road tunnel electromechanical device damage to the system function is only influence on the system integrity, the ew=2, when the influence of the road tunnel electromechanical device damage to the system function is slight influence but the main function is perfect, the ew=3, when the influence of the road tunnel electromechanical device damage to the system function is serious damage to the system function, the ew=5, and when the influence of the road tunnel electromechanical device damage to the system function is invalid, the ew=7;
MW is a cost weight of the highway tunnel electromechanical device, mw=1 when the cost of the highway tunnel electromechanical device is equal to or less than 1 ten thousand yuan, mw=2 when the cost of the highway tunnel electromechanical device is equal to or less than 5 ten thousand yuan, mw=3 when the cost of the highway tunnel electromechanical device is equal to or less than 10 ten thousand yuan, mw=4 when the cost of the highway tunnel electromechanical device is equal to or less than 20 ten thousand yuan, and mw=5 when the cost of the highway tunnel electromechanical device is equal to or more than 20 ten thousand yuan;
step 1.5, determining the value of the number ks of the optimal cluster clusters through an elbow method according to the reliability evaluation RE of the equipment, the maintenance evaluation WE of the equipment and the loss cost QE of the equipment; the index of the elbow method is error square sum SSE, and when inflection points appear in an image of which SSE is reduced along with the increase of the ks value, the corresponding ks value is the number of clusters of the clusters;
wherein: c (C) kk Is a divided cluster, and the value of kk is in a closed interval [1, ks ]]Integers between, xp hg Is C kk Sample points in (us) kk Is C kk The mean value of all samples;
step 1.6, after determining the value of ks, determining each point xpp in the sample by the Euclidean distance dist through the k-means algorithm z Relative to the cluster center C kk Selecting the cluster closest to the distance as the category to which the cluster belongs; then, according to the classified categories and data, calculating the category center, namely the centroid of the cluster again, and repeating the steps until the cluster center is not changed; wherein:
Step 1.7, classifying the levels of the electromechanical equipment by using a CART algorithm; the method comprises the following specific steps:
step 1.7.1, taking a possible value a for a subdivision index A in an electromechanical device index data set D, and dividing the D into two subsets D1 and D2;
step 1.7.2, respectively calculating Gini coefficients of 3 types of subdivision indexes in the data set D after the subset division, and selecting the index with the smallest value as the optimal division index;
step 1.7.3, repeating the first two steps until the number of samples contained in the subset is too small or the Gini coefficient is less than a threshold value;
step 1.7.4, judging the category of the subset according to the mode of the equipment category of the sample in each subset, and generating an electromechanical equipment subdivision decision tree;
1.7.5, pruning subtrees with different sizes according to the cross verification error and the complexity; finally, obtaining an optimal electromechanical device importance degree division decision tree;
for the sample D, a certain value a of the feature a divides D into two parts D1 and D2, and under the condition of the feature a, the Gini coefficient calculation formula is as follows:
pro bn representing the probability that the sample points in D belong to the bn class, uc represents the number of sample classes in D.
Example 2:
the embodiment also provides a construction method of the residual life prediction model of the road tunnel electromechanical equipment, which comprises the following steps:
Step 2.1, acquiring a life characteristic data set of the electromechanical equipment of the highway tunnel:
collecting internal state parameter data and external environment parameter data of the road tunnel electromechanical equipment, establishing an original data set, cleaning and denoising the original data set, grouping cleaned data according to a fixed time length step length, and updating the data set to obtain an electromechanical equipment life characteristic data set;
the internal state parameters of the road tunnel electromechanical equipment are voltage, current, power, vibration, maintenance times, composition complexity and operation duration;
the external environment parameters are temperature, humidity, wind power, wind speed, extreme climate times, CO concentration, nitrogen oxide concentration and PM value;
in step 2.1, the specific acquisition process of the life characteristic data set of the road tunnel electromechanical equipment is as follows:
step S2.1.1, determining key equipment in the highway tunnel according to the highway maintenance engineering management method of the transportation department, and taking the key equipment as highway tunnel electromechanical equipment for predicting the residual life;
the key equipment needs to meet the following two conditions: the running state of the equipment has important influence on driving safety or operation management in the tunnel; the second condition is that the fault burst rate of the key equipment is higher, and the main equipment with little long-term change of the equipment state in the tunnel is eliminated;
S2.1.2, determining internal state parameters of road tunnel key equipment, including voltage, current, power, vibration and operation time length, according to the road engineering electromechanical facility standard assembly, wherein state parameter indexes can reflect the operation state of the electromechanical equipment and provide input data support for a prediction model;
step S2.1.3, collecting external environmental parameters of the electromechanical device, including temperature, humidity and wind power data;
step S2.1.4, establishing an electromechanical device state parameter acquisition module aiming at the state parameters proposed in step S2.1.2, acquiring original monitoring data and storing historical monitoring data by adopting different working conditions to obtain characteristic parameter data of each key device under different working conditions, and combining the external environment parameter data of the device to form an original data set;
and S2.1.5, cleaning and denoising the original data set, adopting a time sequence data processing method to perform dimensionless and vector processing on the acquired running state data of the tunnel electromechanical equipment, and updating the EQLC data set.
Step 2.2, determining the failure rate of the electromechanical equipment of the highway tunnel:
establishing a Weibull distribution fault model which accords with the failure rule of the road tunnel electromechanical equipment, and determining the failure rate of the road tunnel electromechanical equipment through the Weibull distribution fault model;
The Weibull distribution fault model not only can represent the fault characteristics of the electromechanical equipment, but also can convert the operation state duration of the equipment into the fault rate as a main characteristic parameter.
In step 2.2, the specific determination process of the failure rate of the electromechanical equipment of the highway tunnel is as follows:
and S2.2.1, according to the highway engineering quality identification and assessment standard and the highway maintenance contract, the maintenance stage of the highway is mainly developed into a preventive maintenance stage. Therefore, the improved Weibull distribution curve model can be adopted for evaluation, the improved Weibull distribution defect period, the maintenance period and the distribution function of the equipment model are as follows:
wherein t is time, alpha is a shape parameter, and eta is a scale parameter;
in step S2.2.2, deriving t from the distribution function, a probability density function is obtained as follows:
step S2.2.3, the failure rate function of the electromechanical device conforming to the highway tunnel is as follows:
wherein R is a reliability function.
Step S2.2.4, estimating the parameters of the Weibull model by using a least square estimation method according to the historical data of the electromechanical equipment to obtain a fault model capable of representing the fault characteristics of the electromechanical equipment;
and S2.2.5, converting the operation state duration of the equipment into a fault rate by using the electromechanical equipment fault model, and qualitatively analyzing the approximate development trend of the service life of the electromechanical equipment. And taking the failure rate as a main characteristic parameter of the pre-residual life of the electromechanical equipment, and constructing a life prediction characteristic vector of the electromechanical equipment of the highway tunnel by combining other parameters.
Step 2.3, compensation of internal state parameters of the road tunnel electromechanical device:
an external environment parameter is adopted to establish an environment factor error identification model, and the internal state parameter of the road tunnel electromechanical equipment is compensated to obtain the compensated internal state parameter;
by analyzing the structure and the construction mechanism of the key equipment, the reference influence relation of the environmental factors on the characteristic parameters of the electromechanical equipment can be obtained, the characteristic parameters which are not influenced by the environmental factors in the electromechanical equipment are deleted in theory, and the characteristic parameters do not need to be compensated by the environmental factors;
for the key electromechanical equipment with an incompletely clear structure and incompletely clear factors, the correlation between the characteristic parameters of the equipment and the environmental factors is incompletely clear, and the gray correlation model is required to analyze all parameter data so as to measure the influence of the environmental factors on the test parameter data;
the basic idea of the gray correlation model is to convert the discrete value of each array in the system into a continuous curve, judge whether the connection is tight by comparing the similarity degree of continuous geometric curves of the sequences, and the higher the similarity degree of the geometric curves is, the larger the correlation degree between the corresponding sequences is, and otherwise, the smaller the correlation degree between the corresponding sequences is;
In step 2.3, the specific process of compensating the internal state parameters of the electromechanical equipment of the highway tunnel is as follows, namely, the characteristic parameter data of the electromechanical equipment is subjected to characteristic analysis of environmental factors by adopting the gray correlation model. The method comprises the following specific steps:
step S2.3.1, selecting a key device, determining specific environmental factors, and carrying out normalization processing and initialization transformation on each characteristic parameter of the key device;
step S2.3.2, respectively calculating gray correlation degrees of all the test data relative to the environmental factors by taking the determined environmental factor data as a reference sequence;
step S2.3.3, sorting gray relevancy according to the order of magnitude, finding out electromechanical device characteristic parameters of relevancy threshold, and describing that environmental factors have great influence on the parameters, wherein in order to extract electromechanical device residual life characteristic factors and perform performance evaluation and prediction research for a long time by utilizing the characteristic parameter data, the parameters need to be further subjected to environmental factor compensation and modeling;
step S2.3.4, for the above-mentioned electromechanical device characteristic parameters, analyzing according to the environmental factors and the measured parameter data, using a unitary linear regression model to perform least square fitting on the characteristic parameters and the environmental factors, and establishing an environmental factor error identification model, thereby effectively eliminating the influence of the environmental factors on the electromechanical device characteristic parameters.
Step 2.4, constructing life characteristic vectors of the electromechanical equipment of the highway tunnel:
taking the compensated internal state parameters and the fault rate of the road tunnel electromechanical equipment as analysis objects, and screening the analysis objects by adopting an unsupervised feature extraction method of nuclear principal component analysis to obtain feature vectors capable of representing the service life of the road tunnel electromechanical equipment;
in step 2.4, the specific construction process of the life characteristic vector of the road tunnel electromechanical equipment is as follows:
because the monitoring parameters of the road tunnel electromechanical equipment are more and nonlinear correlation exists among the parameter monitoring values, in order to eliminate the nonlinear correlation among the historical state parameter data, the data dimension is reduced, and the feature parameters subjected to the environmental factor compensation treatment are effectively screened by adopting an unsupervised feature extraction method of Kernel Principal Component Analysis (KPCA). The key point of the KPCA method is that a nonlinear mapping function is utilized to map a data set with relevance into a high-dimensional feature space, then traditional principal component analysis is carried out, and a kernel matrix is used for replacing an inner product matrix in the high-dimensional feature space;
the analysis and calculation flow of the core principal component is as follows:
step S2.4.1, setting b as the total number of the road tunnel electromechanical equipment monitoring data samples, and m as the number of the characteristic parameters, selecting a nuclear function meeting the requirements through analysis, and obtaining a corresponding nuclear matrix H according to the nuclear function and a sample matrix;
In step S2.4.2, the core matrix is further centered, and the method for centering the matrix H is as follows:
wherein I is a matrix with the value of b multiplied by b being 1,is a processed nuclear matrix;
step S2.4.3, findCharacteristic value theta of ω Corresponding feature vector L ω (ω＝1,2,...,b)；
Step S2.4.4, finding the kernel principal component vector l of phi ω The following formula is shown:
wherein phi is a sample matrix mapped to a high-dimensional space;
step S2.4.5, calculating a variance contribution rate and a cumulative contribution rate:
wherein θ ω Variance of principal component omega; zeta type toy ω Variance contribution rate for principal component omega; η (eta) g The variance contribution rate is accumulated for the g-th principal component.
Step 2.5, establishing a life prediction model of the road tunnel electromechanical equipment based on the cyclic neural network:
the circulating neural network is a long-period memory network, the characteristic vector is used as the input quantity of the long-period memory network, the residual life of the road tunnel electromechanical equipment is used as the output quantity of the long-period memory network, and the long-period memory network is trained to obtain a life prediction model of the road tunnel electromechanical equipment;
in step 2.5, the specific establishment process of the life prediction model of the road tunnel electromechanical equipment based on the cyclic neural network is as follows:
And predicting the life of the road tunnel electromechanical equipment by using a Long Short-Term Memory (LSTM) network. LSTM is an excellent improved algorithm based on a cyclic neural network (RNN), inherits the advantage that the cyclic neural network can better process time series data, and solves the problem that gradient explosion or disappearance occurs when the time series are unfolded for a long time. The procedure for constructing LSTM-based electromechanical device lifetime prediction is as follows;
step S2.5.1, initializing a network, and respectively assigning a random number between (0 and 1) to each connection weight and bias;
step S2.5.2, using the public factor matrix F as the input of the network, and using the residual life of the electromechanical equipment as the output;
step S2.5.3, determining the structure of the LSTM neural network, such as the number of LSTM layers and the number of neurons in each layer;
and S2.5.4, training the LSTM neural network by using a training set, and if the error requirement is met or the maximum iteration number is exceeded, finishing model training to obtain the life prediction model of the road tunnel electromechanical equipment.
The embodiment also provides a method for predicting the residual life of the road tunnel electromechanical device, which adopts a life prediction model of the road tunnel electromechanical device to predict the life of the road tunnel electromechanical device, wherein the input quantity of the life prediction model of the road tunnel electromechanical device is a characteristic vector obtained according to the method from step 2.1 to step 2.4, and the output quantity of the life prediction model of the road tunnel electromechanical device is the residual life of the road tunnel electromechanical device.
According to the invention, equipment state information obtained in real time in the operation period of the tunnel electromechanical system is used as a basis, the operation state duration of equipment is converted into failure rate to serve as a main characteristic parameter by constructing a failure model representing the life distribution characteristics of the electromechanical equipment, and a life prediction characteristic vector of the electromechanical equipment is constructed by combining other related characteristic parameters, so that environment modeling compensation and data preprocessing are further carried out on long-term monitoring data of the electromechanical equipment, characteristic factors representing the life of the electromechanical equipment are extracted to serve as input of a prediction model, the residual life of the electromechanical equipment serves as output of the prediction model, and a life prediction model of the highway tunnel electromechanical equipment based on a cyclic neural network is constructed.
Electromechanical device lifetime characteristics (Electromechanical Equipment Life Characteristics, EQLC for short).
Long Short-Term Memory (LSTM) networks.
Example 3:
the embodiment provides a method for constructing a degradation analysis model of electromechanical equipment of a highway, which comprises the following steps:
step 3.1, obtaining original data:
the original data comprise synchronous operation state monitoring data, health state monitoring data, operation environment condition monitoring data and historical maintenance data of the highway electromechanical equipment;
The operation state monitoring data comprise voltage, current and power factor of the highway electromechanical equipment;
the health state monitoring data comprise illumination light flux maintenance rate, storage battery capacity or PLC read-write speed;
the operation environment condition monitoring data comprise the environment temperature, the environment humidity and vibration monitoring values of the highway electromechanical equipment;
the historical maintenance data comprise maintenance time and maintenance work grade of the highway electromechanical equipment in the same period;
step 3.2, constructing an operation state characteristic parameter sequence and a health state monitoring data sequence:
setting a unit time interval as t, and setting operation state characteristic parameter sets of the electromechanical equipment of the highway in a plurality of unit time intervals: [ v_mean, v_std, c_mean, c_std, p_mean, p_std ], each column of the running state characteristic parameter set within a plurality of unit time intervals forms a running state characteristic parameter sequence;
wherein: v represents voltage, c represents current, p represents power factor, mean represents average value of sampling values in unit time interval t, std represents standard deviation of sampling values in unit time interval t;
taking the average value of sampling values of the health state monitoring data in a time interval t as an equipment health index parameter HI, and forming a health state monitoring data sequence by the health state monitoring data of the highway electromechanical equipment in a plurality of unit time intervals which correspond to the running state characteristic parameter sequence;
Step 3.3, screening sensitive parameters:
calculating a gray correlation value R by adopting a gray correlation entropy algorithm, taking an operation state characteristic parameter sequence as a comparison sequence and a health state monitoring data sequence as a reference sequence;
calculating Pearson correlation coefficients r of the running state characteristic parameter sequence and the health state monitoring data sequence by adopting a correlation analysis method;
calculating a weighted average value according to y=0.4|R|+0.6|r|, and arranging six parameters of the running state characteristic parameter set in a descending order according to y, and selecting the first three parameters as sensitive parameters;
step 3.4, constructing a cumulative effect parameter sequence influenced by external factors:
setting a factor level of the operating environment condition monitoring data:
the level of 1 factor of the ambient temperature is [15,18 ], the level of 2 factor of the ambient temperature is [18,21 ], the level of 3 factor of the ambient temperature is [21,24 ], the level of 4 factor of the ambient temperature is [24, 27), and the level of 5 factor of the ambient temperature is [27, 30);
the level of factor 1 of the ambient humidity is [65%, 70%), the level of factor 2 of the ambient humidity is [70%, 75%), the level of factor 3 of the ambient humidity is [75%, 80%), the level of factor 4 of the ambient humidity is [80%, 85%), the level of factor 5 of the ambient humidity is [85%, 90%);
The level 1 factor level of the vibration monitoring value is [7,8 ] Hz, the level 2 factor level of the vibration monitoring value is [8,9 ] Hz, the level 3 factor level of the vibration monitoring value is [9,10 ] Hz, the level 4 factor level of the vibration monitoring value is [10,11 ] Hz, and the level 5 factor level of the vibration monitoring value is [11,12 ] Hz;
taking the weighted sum of the factor level values of the running environment condition monitoring data taking the occurrence frequency of each level of factor level as a weight value as parameters T_ws, H_ws and V_ws in a period before each time point;
taking standard deviation of a factor level value change sequence of each operation environment condition monitoring data as parameters T_std, H_std and V_std in a period before each time point;
taking the number of days of the last maintenance of the electromechanical equipment of the expressway from each time point as a parameter M_days, wherein the equipment is subjected to the one-time maintenance at the initial moment;
taking the weighted sum of the frequency as a weight value of the level of maintenance work performed on the equipment by the highway electromechanical equipment before each time point as a parameter M_ws;
the accumulated effect parameter set influenced by external factors of the electromechanical equipment of the highway in a plurality of unit time intervals corresponding to the characteristic parameter sequence of the running state is as follows:
[T_ws，H_ws，V_ws，T_std，H_std，V_std，M_dats，M_ws]；
each column of data in the cumulative effect parameter set is formed by external factors in a plurality of unit time intervals to form a cumulative effect parameter sequence;
Wherein: t is a temperature sampling value, H is a humidity sampling value, V is a vibration frequency sampling value, ws is a weighted sum, and std is a standard deviation;
step 3.5, screening the main external cumulative effect influence parameters:
calculating Pearson correlation coefficients r of the accumulated effect parameter sequences and the health state monitoring data sequences influenced by external factors by adopting a correlation analysis method; and according to the r value, the eight parameters affecting the row direction in the cumulative effect parameter set for the external factors are arranged in a descending order. Taking the first three parameters as main external cumulative effect influencing parameters of the degradation process;
step 3.6, establishing a degradation analysis model of the electromechanical equipment of the expressway:
selecting a transformation parameter lambda of a Box-Cox transformation family in a Box-Cox algorithm by adopting a maximum likelihood estimation method, and transforming the health state monitoring data by adopting the Box-Cox algorithm to obtain transformed health state monitoring data;
and constructing a training data set by adopting the converted health state monitoring data, the sensitive parameters and the main external cumulative effect influence parameters, adopting a Bayesian learning framework to minimize the regular error, training a related vector machine model, and obtaining a degradation analysis model of the electromechanical equipment of the expressway.
Example 4:
the embodiment provides a life prediction system of highway electromechanical equipment facing to a full life cycle, wherein the highway electromechanical equipment at least comprises highway tunnel electromechanical equipment and expressway electromechanical equipment, the system comprises a residual life prediction module of the highway tunnel electromechanical equipment, and the residual life prediction module of the highway tunnel electromechanical equipment is used for predicting the residual life of the highway tunnel electromechanical equipment; the highway tunnel electromechanical device residual life prediction module comprises a highway tunnel electromechanical device residual life prediction model, and the highway tunnel electromechanical device residual life prediction model is constructed by adopting the highway tunnel electromechanical device residual life prediction model construction method described in the embodiment 2.
The system also comprises a highway tunnel key electromechanical equipment identification module, wherein the highway tunnel key electromechanical equipment identification module is used for identifying and screening key equipment in highway tunnel electromechanical equipment; the residual life prediction module of the road tunnel electromechanical equipment is used for predicting the residual life of key equipment identified and screened from the road tunnel electromechanical equipment. The highway tunnel key electromechanical device identification module described in embodiment 1 is used.
The system also comprises a degradation analysis module of the expressway electromechanical equipment, wherein the degradation analysis module of the expressway electromechanical equipment is used for carrying out degradation analysis on the expressway electromechanical equipment; the degradation analysis module of the highway electromechanical equipment comprises a degradation analysis model of the highway electromechanical equipment, and the degradation analysis model of the highway electromechanical equipment is constructed by adopting the degradation analysis model construction method of the highway electromechanical equipment in the embodiment 3.
Application example 1:
the above embodiment 4 is followed, the PLC control box of the highway tunnel power supply and distribution system is determined as a key device by the highway tunnel key electromechanical device identification module described in embodiment 1, and the method for predicting the remaining life of the highway tunnel electromechanical device in embodiment 2 is followed, and the remaining life is predicted using the monitoring data. The method comprises the steps of determining internal state parameters of a PLC control box to be voltage, current, power, vibration and operation duration, collecting external environment parameters affecting the operation of the control box to be temperature, humidity and wind power, and constructing an EQLC data set of the PLC control box, wherein an original record comprises 10 fields, specifically equipment number, EQLC number, operation duration, voltage, current, power, vibration, temperature, humidity and wind power. And converting the operation time length parameter into a fault rate according to the second step, and replacing the operation time length attribute in the EQLC. And further compensating the internal parameters of the electromechanical equipment, taking the compensated internal state parameters and failure rate as analysis objects, screening by utilizing core principal component analysis, obtaining characteristic vectors of the service life of the PLC control box as input data of the LSTM prediction network, and inputting the data into the trained prediction network to obtain the predicted residual service life.
As shown in fig. 5, the predicted expected residual life values of the method were obtained at 50% and 80% of the sites of the PLC control box, and the specific results are shown in table 1 below.
TABLE 1 residual service life at different quantiles
Application example 2:
following the above-described embodiment 4, the degradation analysis model of the highway electromechanical device in this application example is obtained by taking the UPS power supply device battery pack as an example, according to the construction method of the degradation analysis model of the highway electromechanical device as described in embodiment 3.
Wherein:
in step 3.1, the operating state monitoring data sample of the storage battery pack of the UPS power supply apparatus is shown in table 2, the operating environment condition monitoring data sample is shown in table 3, and the history maintenance data sample is shown in table 4.
In step 3.2, the characteristic parameter set of the running state of the battery pack of the UPS power supply apparatus is shown in table 5, and the health status monitoring data sequence of the battery pack of the UPS power supply apparatus is shown in table 6.
In step 3.3, the gray correlation value R and the Pearson correlation coefficient R of the running state characteristic parameter sequence of the storage battery pack of the UPS power supply device are shown in table 7, and the descending order of the six running parameters in the running state characteristic parameter set is shown in table 8.
In step 3.4, the external factors of the highway electromechanical device of the UPS power source device battery pack affect the cumulative effect parameter set samples, for example, as shown in table 9.
In step 3.5, the Pearson correlation r of the cumulative effect parameter sequence is shown in table 10, as influenced by external factors of the battery pack of the UPS power supply apparatus.
In step 3.6, RVM model:wherein:
RVM model is related vector machine model; k (x, x) i ) Is a Gaussian kernel function; x is an input vector formed by sensitive parameters and external factors accumulated to influence effect parameters; x is x i The kernel function is at the center point of the ith iteration; w (w) i Is a model weight; n is the number of samples; w (w) 0 Is the deviation.
The model visualization results are shown in fig. 6, for example.
Table 2 running state monitoring data sample
| Sampling time point | Output voltage (v) | Output current (mA) | Output power factor |
| 1 | 3.974870912 | -2.012528324 | 0.75 |
| 2 | 3.951716708 | -2.013979362 | 0.73 |
| … | … | … | … |
TABLE 3 operating Environment Condition monitoring data sample
| Sampling time point | Temperature (. Degree. C.) | Humidity (%) | Vibration (Hz) |
| 1 | 21.9 | 72.7 | 10.6 |
| 2 | 21.8 | 73.1 | 8.8 |
| … | … | … | … |
Table 4 historical maintenance data sample
| Sequence number | Date of day | Maintenance level |
| 1 | 8 months and 2 days | 2 |
| 2 | 9 months 3 days | 4 |
| … | … | … |
TABLE 5 run State characteristic parameter set sample
| Time interval sequence number | v_mean | v_std | c_mean | c_std | p_mean | p_std |
| 1 | 3.52982 | 0.05567 | -1.81870 | 0.35229 | 0.74 | 0.11325 |
| 2 | 3.53732 | 0.05511 | -1.81755 | 0.35423 | 0.76 | 0.13645 |
| … | … | … | … | … | … | … |
TABLE 6 health status monitoring data sequence samples
TABLE 7 Gray correlation value of operating State characteristic parameter sequence R and Pearson correlation coefficient R
| Parameter name | v_mean | v_std | c_mean | c_std | p_mean | p_std |
| R | 0.92350 | 0.93714 | 0.80098 | 0.87546 | 0.78881 | 0.85951 |
| r | -0.61117 | 0.83858 | 0.68347 | 0.80103 | 0.70508 | -0.74843 |
Table 8 results of descending order of six parameters of the row direction in the running state characteristic parameter set
| v_mean | v_std | c_mean | c_std | p_mean | p_std | |
| y | 0.736102 | 0.878004 | 0.730474 | 0.830802 | 0.738572 | 0.792862 |
| Ordering of | 5 | 1 | 6 | 2 | 4 | 3 |
TABLE 9 external factor influence cumulative Effect parameter set sample
| Sequence number | T_ws | H_ws | V_ws | T_std | H_std | V_std | M_days | M_ws |
| 1 | 2 | 3 | 4 | 1.265 | 0.907 | 2.576 | 30 | 3 |
| 2 | 3 | 2 | 3 | 1.129 | 0.974 | 3.545 | 25 | 4 |
| … | … | … | … | … | … | … | … | … |
TABLE 10 Pearson correlation coefficient r for cumulative effect parameter sequence affected by external factors
| Parameters (parameters) | T_ws | H_ws | V_ws | T_std | H_std | V_std | M_days | M_ws |
| r | 0.74325 | 0.62734 | 0.68347 | 0.80103 | 0.55768 | 0.67428 | 0.32758 | 0.53724 |
Claims (3)
1. The life prediction system for the highway electromechanical equipment facing the full life cycle is characterized by comprising a residual life prediction module of the highway tunnel electromechanical equipment, wherein the residual life prediction module of the highway tunnel electromechanical equipment is used for predicting the residual life of the highway tunnel electromechanical equipment; the highway tunnel electromechanical equipment residual life prediction module comprises a highway tunnel electromechanical equipment residual life prediction model, and the highway tunnel electromechanical equipment residual life prediction model is constructed according to the following method:
step 2.1, acquiring a life characteristic data set of the electromechanical equipment of the highway tunnel:
collecting internal state parameter data and external environment parameter data of the road tunnel electromechanical equipment, establishing an original data set, cleaning and denoising the original data set, grouping cleaned data according to a fixed time length step length, and updating the data set to obtain an electromechanical equipment life characteristic data set;
The internal state parameters of the road tunnel electromechanical equipment are voltage, current, power, vibration, maintenance times, composition complexity and operation duration;
the external environment parameters are temperature, humidity, wind power, wind speed, extreme climate times, CO concentration, nitrogen oxide concentration and PM value;
step 2.2, determining the failure rate of the electromechanical equipment of the highway tunnel:
establishing a Weibull distribution fault model which accords with the failure rule of the road tunnel electromechanical equipment, and determining the failure rate of the road tunnel electromechanical equipment through the Weibull distribution fault model;
step 2.3, compensation of internal state parameters of the road tunnel electromechanical device:
an external environment parameter is adopted to establish an environment factor error identification model, and the internal state parameter of the road tunnel electromechanical equipment is compensated to obtain the compensated internal state parameter;
in step 2.3, the specific steps for compensating the internal state parameters of the road tunnel electromechanical device are as follows:
step S2.3.1, selecting a key device, determining specific environmental factors, and carrying out normalization processing and initialization transformation on each characteristic parameter of the key device;
step S2.3.2, respectively calculating gray correlation degrees of all the test data relative to the environmental factors by taking the determined environmental factor data as a reference sequence;
Step S2.3.3, sorting gray relevancy according to the order of magnitude, finding out electromechanical device characteristic parameters of relevancy threshold, and describing that environmental factors have great influence on the parameters, wherein in order to extract electromechanical device residual life characteristic factors and perform performance evaluation and prediction research for a long time by utilizing the characteristic parameter data, the parameters need to be further subjected to environmental factor compensation and modeling;
s2.3.4, analyzing the characteristic parameters of the electromechanical device according to the environmental factors and the measured parameter data, and performing least square fitting on the characteristic parameters and the environmental factors by adopting a unitary linear regression model to establish an environmental factor error identification model, so that the influence of the environmental factors on the characteristic parameters of the electromechanical device is effectively eliminated;
step 2.4, constructing life characteristic vectors of the electromechanical equipment of the highway tunnel:
taking the compensated internal state parameters and the fault rate of the road tunnel electromechanical equipment as analysis objects, and screening the analysis objects by adopting an unsupervised feature extraction method of nuclear principal component analysis to obtain feature vectors capable of representing the service life of the road tunnel electromechanical equipment;
step 2.5, establishing a life prediction model of the road tunnel electromechanical equipment based on the cyclic neural network:
The circulating neural network is a long-period memory network, the characteristic vector is used as the input quantity of the long-period memory network, the residual life of the road tunnel electromechanical equipment is used as the output quantity of the long-period memory network, and the long-period memory network is trained to obtain the life prediction model of the road tunnel electromechanical equipment.
2. The life-cycle oriented highway electromechanical device life prediction system of claim 1, further comprising a highway tunnel key electromechanical device identification module for identifying and screening key devices in the highway tunnel electromechanical device; the highway tunnel electromechanical equipment residual life prediction module is used for predicting the residual life of key equipment identified and screened from the highway tunnel electromechanical equipment; the highway tunnel key electromechanical equipment identification module comprises the following steps:
step 1.1, acquiring the number gn of faults of the road tunnel electromechanical equipment in the monitored days, the equipment cost em and the maintenance cost wc according to the maintenance record of the road tunnel electromechanical equipment, and recording the maintenance grade drq of the road tunnel electromechanical equipment;
Wherein:
if the road tunnel electromechanical equipment is replaced due to direct damage, the fault frequency gn of the road tunnel electromechanical equipment is still increased by 1;
when the maintenance process is daily inspection work on the appearance and the working state of the equipment, checking whether the working function, the performance and the running environment of the equipment meet the requirements or not and recording potential safety hazards, drq=1;
when the maintenance process is to carry out external periodic maintenance on the equipment; checking whether the shells of various electromechanical equipment leak electricity, and when the equipment is fastened, adjusted, derusted and preserved, drq=2;
when the maintenance process is to clean the inside of the equipment; maintaining the internal element, and fastening the internal connecting wire and the parts; when parts with serious abrasion are maintained and replaced and basic function tests of partial equipment are carried out, drq=3;
when the maintenance process is to perform overall maintenance and overhaul on equipment, including grounding test, core element inspection, fastening detection of wiring and parts, replacement of parts with serious damage or aging, and detection and debugging of main functions of the equipment, drq=4;
when the maintenance process is to comprehensively use technical means to test the key performance of the electromechanical equipment, the service life of the existing equipment, the use perfection rate of the equipment and the fault occurrence rate of the equipment are evaluated; drq=5 when evaluating the function and performance of the device;
Step 1.2, the reliability evaluation RE of the computing device,；
wherein,days for monitoring the device;
step 1.3, calculating the maintenance difficulty DE, maintenance cost and single price ratio BE of equipment; performing standardized treatment on the DE and the BE to obtain the result Dd and Bb, and calculating the maintenance evaluation WE of the equipment;
；/>；/>；
step 1.4, calculating a loss cost QE of the device,wherein:
the EW is an influence weight of the road tunnel electromechanical device damage to the system function, and when the influence of the road tunnel electromechanical device damage to the system function is no influence, the ew=1, when the influence of the road tunnel electromechanical device damage to the system function is only influence on the system integrity, the ew=2, when the influence of the road tunnel electromechanical device damage to the system function is slight influence but the main function is perfect, the ew=3, when the influence of the road tunnel electromechanical device damage to the system function is serious damage to the system function, the ew=5, and when the influence of the road tunnel electromechanical device damage to the system function is invalid, the ew=7;
MW is a cost weight of the highway tunnel electromechanical device, mw=1 when the cost of the highway tunnel electromechanical device is equal to or less than 1 ten thousand yuan, mw=2 when the cost of the highway tunnel electromechanical device is equal to or less than 5 ten thousand yuan, mw=3 when the cost of the highway tunnel electromechanical device is equal to or less than 10 ten thousand yuan, mw=4 when the cost of the highway tunnel electromechanical device is equal to or less than 20 ten thousand yuan, and mw=5 when the cost of the highway tunnel electromechanical device is equal to or more than 20 ten thousand yuan;
Step 1.5, determining the value of the number ks of the optimal cluster clusters through an elbow method according to the reliability evaluation RE of the equipment, the maintenance evaluation WE of the equipment and the loss cost QE of the equipment; the index of the elbow method is error square sum SSE, and when inflection points appear in an image of which SSE is reduced along with the increase of the ks value, the corresponding ks value is the number of clusters of the clusters;
wherein:is a divided cluster, +.>The value is in the closed interval [1, ks ]]Integer between>Is->In (c) the sample points in (c),is->The mean value of all samples;
step 1.6, after the value of ks is determined, each point in the sample is determined through a k-means algorithm and through a Euclidean distance distRelative to the cluster center->Selecting the cluster closest to the distance as the category to which the cluster belongs; then, according to the classified categories and data, calculating the category center, namely the centroid of the cluster again, and repeating the steps until the cluster center is not changed; wherein:
step 1.7, classifying the levels of the electromechanical equipment by using a CART algorithm; the method comprises the following specific steps:
step 1.7.1, taking a possible value a for a subdivision index A in an electromechanical device index data set D, and dividing the D into two subsets D1 and D2;
step 1.7.2, respectively calculating Gini coefficients of 3 types of subdivision indexes in the data set D after the subset division, and selecting the index with the smallest value as the optimal division index;
Step 1.7.3, repeating the first two steps until the number of samples contained in the subset is too small or the Gini coefficient is less than a threshold value;
step 1.7.4, judging the category of the subset according to the mode of the equipment category of the sample in each subset, and generating an electromechanical equipment subdivision decision tree;
1.7.5, pruning subtrees with different sizes according to the cross verification error and the complexity; finally, obtaining an optimal electromechanical device importance degree division decision tree;
for the sample D, a certain value a of the feature a divides D into two parts D1 and D2, and under the condition of the feature a, the Gini coefficient calculation formula is as follows:
indicating that the sample point in D belongs to->Probability of class->Representing the number of sample categories in D.
3. The life cycle oriented highway electromechanical device life prediction system of claim 1, further comprising a highway electromechanical device degradation analysis module for performing degradation analysis on the highway electromechanical device; the degradation analysis module of the highway electromechanical equipment comprises a degradation analysis model of the highway electromechanical equipment, and the degradation analysis model of the highway electromechanical equipment is constructed according to the following method:
Step 3.1, obtaining original data:
the original data comprise synchronous operation state monitoring data, health state monitoring data, operation environment condition monitoring data and historical maintenance data of the highway electromechanical equipment;
the operation state monitoring data comprise voltage, current and power factor of the highway electromechanical equipment;
the health state monitoring data comprise illumination light flux maintenance rate, storage battery capacity or PLC read-write speed;
the operation environment condition monitoring data comprise the environment temperature, the environment humidity and vibration monitoring values of the highway electromechanical equipment;
the historical maintenance data comprise maintenance time and maintenance work grade of the highway electromechanical equipment in the same period;
step 3.2, constructing an operation state characteristic parameter sequence and a health state monitoring data sequence:
setting a unit time interval as t, and setting operation state characteristic parameter sets of the electromechanical equipment of the highway in a plurality of unit time intervals:each column of data in the running state characteristic parameter set in a plurality of unit time intervals forms a running state characteristic parameter sequence;
wherein:indicating voltage, < >>Indicating current,/->Representing the power factor >Represents the mean value of the sampled values in a unit time interval t, < >>Representing the standard deviation of sampling values in a unit time interval t;
taking the average value of sampling values of health state monitoring data in a time interval t as a device health index parameterThe health state monitoring data of the highway electromechanical equipment in a plurality of unit time intervals corresponding to the running state characteristic parameter sequences form a health state monitoring data sequence;
step 3.3, screening sensitive parameters:
calculating gray correlation value by adopting a gray correlation entropy algorithm, taking an operation state characteristic parameter sequence as a comparison sequence and a health state monitoring data sequence as a reference sequence；
Calculating Pearson correlation coefficient of the running state characteristic parameter sequence and the health state monitoring data sequence by adopting a correlation analysis method；
According toCalculate the weighted average and according to +.>The six parameters of the running state characteristic parameter set in the row direction are arranged in a descending order, and the first three parameters are selected as sensitive parameters;
step 3.4, constructing a cumulative effect parameter sequence influenced by external factors:
setting a factor level of the operating environment condition monitoring data:
the level of 1 factor of the ambient temperature is [15,18 ], the level of 2 factor of the ambient temperature is [18,21 ], the level of 3 factor of the ambient temperature is [21,24 ], the level of 4 factor of the ambient temperature is [24, 27), and the level of 5 factor of the ambient temperature is [27, 30);
The level of factor 1 of the ambient humidity is [65%, 70%), the level of factor 2 of the ambient humidity is [70%, 75%), the level of factor 3 of the ambient humidity is [75%, 80%), the level of factor 4 of the ambient humidity is [80%, 85%), the level of factor 5 of the ambient humidity is [85%, 90%);
the level 1 factor level of the vibration monitoring value is [7,8 ] Hz, the level 2 factor level of the vibration monitoring value is [8,9 ] Hz, the level 3 factor level of the vibration monitoring value is [9,10 ] Hz, the level 4 factor level of the vibration monitoring value is [10,11 ] Hz, and the level 5 factor level of the vibration monitoring value is [11,12 ] Hz;
taking the weighted sum of the factor level values of the monitoring data of each running environment condition taking the occurrence frequency of each level of factor level as the weight value as a parameter in the period before each time point；
Taking standard deviation of factor level value change sequences of all the operation environment condition monitoring data as parameters in a period before each time point；
Taking the number of days of last maintenance of the electromechanical equipment of the expressway at each time point as a parameterWherein the equipment is maintained once at the initial time;
taking the weighted sum of frequency as weight as parameter of the level of maintenance work performed on the equipment by the electromechanical equipment of the expressway before each time point ；
The accumulated effect parameter set influenced by external factors of the electromechanical equipment of the highway in a plurality of unit time intervals corresponding to the characteristic parameter sequence of the running state is as follows:
；
each column of data in the cumulative effect parameter set is formed by external factors in a plurality of unit time intervals to form a cumulative effect parameter sequence;
wherein:is the temperature sampling value, < >>Is the humidity sampling value, < >>Is a sampling value of vibration frequency, ">Representing a weighted sum->Representing standard deviation;
step 3.5, screening the main external cumulative effect influence parameters:
calculating Pearson correlation coefficient of external factors affecting the accumulated effect parameter sequence and the health state monitoring data sequence by adopting a correlation analysis methodThe method comprises the steps of carrying out a first treatment on the surface of the And according to->The values are arranged in descending order on eight parameters affecting the row direction in the cumulative effect parameter set by external factors, and the first three parameters are taken as main external cumulative effect affecting parameters in the degradation process;
step 3.6, establishing a degradation analysis model of the electromechanical equipment of the expressway:
selecting transformation parameters of Box-Cox transformation family in Box-Cox algorithm by using maximum likelihood estimation methodTransforming the health state monitoring data by using a Box-Cox algorithm to obtain transformed health state monitoring data;
And constructing a training data set by adopting the converted health state monitoring data, the sensitive parameters and the main external cumulative effect influence parameters, adopting a Bayesian learning framework to minimize the regular error, training a related vector machine model, and obtaining a degradation analysis model of the electromechanical equipment of the expressway.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010121280.7A CN111382542B (en) | 2020-02-26 | 2020-02-26 | Highway electromechanical device life prediction system facing full life cycle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010121280.7A CN111382542B (en) | 2020-02-26 | 2020-02-26 | Highway electromechanical device life prediction system facing full life cycle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111382542A CN111382542A (en) | 2020-07-07 |
| CN111382542B true CN111382542B (en) | 2024-02-02 |
Family
ID=71215218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010121280.7A Active CN111382542B (en) | 2020-02-26 | 2020-02-26 | Highway electromechanical device life prediction system facing full life cycle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111382542B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111951418B (en) * | 2020-07-28 | 2022-03-04 | 重庆首讯科技股份有限公司 | ETC portal system and prediction method of stable operation time length thereof |
| CN112580875B (en) * | 2020-12-21 | 2022-06-24 | 泉州装备制造研究所 | Fault prediction method and system for power distribution device |
| CN112884210B (en) * | 2021-01-31 | 2022-03-15 | 中国人民解放军63963部队 | Vehicle health management system based on fuzzy clustering |
| CN113779882A (en) * | 2021-09-10 | 2021-12-10 | 中国石油大学(北京) | Method, device, equipment and storage medium for predicting residual service life of equipment |
| CN113657693B (en) * | 2021-10-20 | 2022-02-08 | 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） | Predictive maintenance system and method for intelligent manufacturing equipment |
| CN114781278B (en) * | 2022-06-17 | 2022-09-27 | 天津理工大学 | Electromechanical equipment service life prediction method and system based on data driving |
| CN114819415B (en) * | 2022-06-27 | 2022-09-20 | 中国标准化研究院 | Power equipment fault prediction system based on data analysis |
| CN116401525B (en) * | 2023-02-23 | 2023-09-29 | 兰州工业学院 | Bridge tunneling prediction maintenance method and system based on intelligent induction |
| CN116258483B (en) * | 2023-05-16 | 2023-07-21 | 交通运输部公路科学研究所 | Highway electromechanical equipment running state estimation modeling method based on dynamic diagram |
| CN116861797B (en) * | 2023-08-30 | 2023-11-24 | 湖南华菱线缆股份有限公司 | Tunnel cable residual life prediction method and device based on machine learning |
| CN117195742A (en) * | 2023-10-11 | 2023-12-08 | 深圳市新红景科技开发有限公司 | Circuit experiment board reliability prediction method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016029590A1 (en) * | 2014-08-28 | 2016-03-03 | 北京交通大学 | Fault prediction and condition-based maintenance method for urban rail train bogie |
| CN109726517A (en) * | 2019-01-31 | 2019-05-07 | 西安理工大学 | A kind of equipment method for predicting residual useful life based on multivariable associated data |
| WO2019201176A1 (en) * | 2018-04-17 | 2019-10-24 | 江苏必得科技股份有限公司 | Method and device for predicting crack damage of train component |
-
2020
- 2020-02-26 CN CN202010121280.7A patent/CN111382542B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016029590A1 (en) * | 2014-08-28 | 2016-03-03 | 北京交通大学 | Fault prediction and condition-based maintenance method for urban rail train bogie |
| WO2019201176A1 (en) * | 2018-04-17 | 2019-10-24 | 江苏必得科技股份有限公司 | Method and device for predicting crack damage of train component |
| CN109726517A (en) * | 2019-01-31 | 2019-05-07 | 西安理工大学 | A kind of equipment method for predicting residual useful life based on multivariable associated data |
Non-Patent Citations (1)
| Title |
|---|
| 基于Wiener过程的功率变换器剩余寿命评估方法;邵力为;王友仁;孙权;;机械制造与自动化(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111382542A (en) | 2020-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111382542B (en) | Highway electromechanical device life prediction system facing full life cycle | |
| CN112101220B (en) | Rolling bearing service life prediction method based on unsupervised model parameter migration | |
| CN111325403B (en) | Method for predicting residual life of electromechanical equipment of highway tunnel | |
| Sun et al. | Stacked denoising autoencoder with density-grid based clustering method for detecting outlier of wind turbine components | |
| CN114218872B (en) | DBN-LSTM semi-supervised joint model-based residual service life prediction method | |
| CN116448419A (en) | Zero sample bearing fault diagnosis method based on depth model high-dimensional parameter multi-target efficient optimization | |
| CN114004137A (en) | Multi-source meteorological data fusion and pretreatment method | |
| CN114297918A (en) | Aero-engine residual life prediction method based on full-attention depth network and dynamic ensemble learning | |
| CN112507479B (en) | Oil drilling machine health state assessment method based on manifold learning and softmax | |
| CN110879377A (en) | Metering device fault tracing method based on deep belief network | |
| CN116579768B (en) | Power plant on-line instrument operation and maintenance management method and system | |
| CN112132210A (en) | Electricity stealing probability early warning analysis method based on customer electricity consumption behavior | |
| Mustafa et al. | Fault identification for photovoltaic systems using a multi-output deep learning approach | |
| CN114791993B (en) | Power battery pack SOH prediction method and system | |
| CN115542429A (en) | XGboost-based ozone quality prediction method and system | |
| CN115860177A (en) | Photovoltaic power generation power prediction method based on combined machine learning model and application thereof | |
| CN115618732A (en) | Nuclear reactor digital twin key parameter autonomous optimization data inversion method | |
| Fu et al. | MCA-DTCN: A novel dual-task temporal convolutional network with multi-channel attention for first prediction time detection and remaining useful life prediction | |
| CN117113086A (en) | Energy storage unit load prediction method, system, electronic equipment and medium | |
| Huang et al. | A hybrid bayesian deep learning model for remaining useful life prognostics and uncertainty quantification | |
| Bond et al. | A hybrid learning approach to prognostics and health management applied to military ground vehicles using time-series and maintenance event data | |
| CN116186633A (en) | Power consumption abnormality diagnosis method and system based on small sample learning | |
| CN113821401A (en) | WT-GA-GRU model-based cloud server fault diagnosis method | |
| CN110837932A (en) | Thermal power prediction method of solar heat collection system based on DBN-GA model | |
| CN111428344B (en) | Construction method of degradation analysis model of electromechanical equipment of expressway |
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 |