CN117407757A - Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD - Google Patents
Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD Download PDFInfo
- Publication number
- CN117407757A CN117407757A CN202310259582.4A CN202310259582A CN117407757A CN 117407757 A CN117407757 A CN 117407757A CN 202310259582 A CN202310259582 A CN 202310259582A CN 117407757 A CN117407757 A CN 117407757A
- Authority
- CN
- China
- Prior art keywords
- data
- lifting appliance
- fault diagnosis
- sample
- svdd
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 37
- 238000012360 testing method Methods 0.000 claims abstract description 28
- 238000000605 extraction Methods 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims abstract description 9
- 230000009467 reduction Effects 0.000 claims abstract description 6
- 238000012549 training Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 12
- 230000003993 interaction Effects 0.000 claims description 10
- 238000000513 principal component analysis Methods 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 10
- 230000005540 biological transmission Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 7
- 230000004069 differentiation Effects 0.000 claims description 6
- 238000012423 maintenance Methods 0.000 claims description 4
- 241000197727 Euscorpius alpha Species 0.000 claims description 3
- 108010074864 Factor XI Proteins 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000012217 deletion Methods 0.000 claims description 2
- 230000037430 deletion Effects 0.000 claims description 2
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000004441 surface measurement 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
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B62—LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
- B62D—MOTOR VEHICLES; TRAILERS
- B62D65/00—Designing, manufacturing, e.g. assembling, facilitating disassembly, or structurally modifying motor vehicles or trailers, not otherwise provided for
- B62D65/02—Joining sub-units or components to, or positioning sub-units or components with respect to, body shell or other sub-units or components
- B62D65/18—Transportation, conveyor or haulage systems specially adapted for motor vehicle or trailer assembly lines
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R19/00—Arrangements for measuring currents or voltages or for indicating presence or sign thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
Abstract
The invention discloses an intelligent fault diagnosis method and system for a production line lifting appliance based on PCA-SVDD, and aims to rapidly and accurately diagnose a short circuit fault of the lifting appliance of an automobile production line. The invention firstly collects the voltage data of the lifting appliance and uses a sliding time window to sample the collected data, secondly pre-processes the data and inputs the data into a PCA model to perform dimension reduction and feature extraction, and finally inputs the extracted features into a SVDD model to perform fault diagnosis and judge the current state of the lifting appliance. Through test, the invention can realize rapid and accurate positioning of the fault lifting appliance, reduce the missing report rate and the false report rate, reduce the shutdown loss, have higher sensitivity, robustness and self-adaptive capacity, and can provide scientific basis for the decision of staff.
Description
Technical Field
The invention belongs to the field of fault diagnosis, and relates to an intelligent fault diagnosis method and system for a production line lifting appliance based on PCA-SVDD.
Background
Automobile production lines serve as "lifelines" in the automobile production and manufacturing industry, playing a vital role in the high-speed development of the automobile industry. With the development of scientific technology, the production line is developing towards high intellectualization, automation and integration, more and more large-scale mechanized lifting appliances are introduced into the same production line, the production order relation among devices is obviously enhanced, and the production process is increasingly complicated. As such, during the operation of the production line, the possibility of failure of the whole system increases, and slight failure during the production process may cause irreparable loss, so how to efficiently and accurately perform failure diagnosis on key equipment such as the lifting appliance of the production line has become one of the main problems in the research of the current automobile manufacturing system.
The most common fault type of the lifting appliance of the automobile production line is short circuit, and most of traditional fault diagnosis work is completed by professional technicians and diagnostic experts in related fields, and the degree of reliability is greatly dependent on the prior experience and expertise. However, the existing production equipment is generally complex and has high automation degree, so that the data volume to be analyzed is huge. Such huge data volumes are obviously not achievable by means of analysis by technicians and diagnostic experts, so that increasing the degree of automation and intelligence of the fault diagnosis of the line spreader has become increasingly common to enterprises.
Since voltage data changes when a short circuit fault occurs in a lifting appliance on an automobile production line, a few researches have been conducted at present to diagnose the fault by using a voltage analysis method so as to identify the approximate position of the fault equipment. However, due to the increase of the number of the lifting appliances on the production line, the reinforcement of the association relation among the devices, the weak anti-interference capability of the voltage signals, the large voltage data volume, the short transmission interval and other problems, the difficulty of extracting typical characteristics from the voltage data and carrying out fault diagnosis is increased, and the difficulty is increased,
at present, a plurality of fault diagnosis and positioning methods are proposed in the academy and mainly divided into three modes of diagnosis based on a traditional quantitative model, diagnosis based on a qualitative model and diagnosis based on equipment big data. The fault diagnosis mode based on the traditional quantitative modeling is mainly divided into two modes of state evaluation and parameter evaluation, but most of real life control systems are nondeterministic, so that the difficulty of constructing a model capable of accurately judging faults is greatly increased, the simplest fault diagnosis mode based on state calculation in reality is not adopted, and when the evaluation workload is too large, the parameter evaluation is difficult to realize. The system fault diagnosis method based on the qualitative model also has the defects that the system fault diagnosis method is only based on the system surface measurement, the interaction relation among all physical quantities in the system cannot be identified, the system is easy to be interfered by the subjective knowledge of researchers, and the like, and the detection conclusion is always biased to be relatively conservative and the tiny problem is easy to be ignored. The method based on the equipment big data solves the defects to a certain extent, but also has the problems of complex calculation process and low calculation accuracy.
Disclosure of Invention
In order to solve the problems of the prior art, the invention aims to provide an intelligent fault diagnosis method for a production line lifting appliance based on PCA-SVDD, which aims to rapidly and accurately diagnose the short circuit fault of the lifting appliance on the automobile production line and further improve the safety and stability of the operation of the production line.
The method is realized by the following technical means:
the invention provides an intelligent fault diagnosis method for a production line lifting appliance based on PCA-SVDD, which comprises the following steps:
step 1: sampling original voltage data of the equipment, wherein the acquisition period is T s Three-phase electricity of a, B and C is collected 1 time in one period of time for 0.001 s;
step 2: sampling data by adopting a sliding time window, and selecting the shape of the sliding time window as (100, 1) and the step length as 100 through algorithm test;
step 3: preprocessing the acquired data, finding that the acquired voltage is in a (-5V, 5V) interval, mapping the voltage value to a (0, 65535) interval for convenience of acquisition, setting f (x) as a corresponding voltage value, and setting x as a specific numerical value, wherein the method comprises the following steps:
step 4: the method comprises the steps of performing dimensionality reduction on data by adopting a Principal Component Analysis (PCA), and firstly constructing 300-dimensional random vectors and X= (X) when processing 1 ，X 2 ，……，X n ，……，X 300 ) (n is more than or equal to 1 and less than or equal to 300), sample data are selected, a sample matrix is constructed, and the following transformation is carried out: (1. Ltoreq.i, j. Ltoreq.n)
From the above variations, a normalized matrix a is obtained, and a correlation coefficient matrix R of the normalized matrix is obtained:
R＝[r ij ] p X p ＝A T A/(n-1) (4)
r ij ＝∑a ij ·a ij /(n-1) (5)
solving a characteristic root lambda of a characteristic equation of the matrix R, and determining a main component:
|R-λI p |＝0 (6)
for each lambda j J=1, 2, m, solving the equation rb=λ j b, obtaining unit feature vectorConverting the standard variable into a main component:
formula (8) wherein F 1 Is the 1 st main component, F n Is the nth principal component and is used as the main component,
step 5: the first 17 main components are selected by using Python programming, and the extracted components are input into an SVDD model for fault diagnosis, and the operation result of the SVDD model is as follows:
the SVDD model can be summarized simply as:
the constraint conditions are: (h) i -A) T (h i -A)≤R 2 ，i＝1，2，…，n (9)
In the formula (9), R represents the radius of the hypersphere, m represents the sphere center of the hypersphere, and n represents the number of sample data during training. Need to solve for the inclusion of all data h i Minimum hyper-sphere of (2)A body.
The SVDD model searching hypersphere can be summarized as a parameter iteration optimization problem, and the basic form is as follows:
and in the constraint condition, introducing a non-negative parameter relaxation factor xi and a penalty factor C.
After introducing the relaxation factor, the minimum problem of the hypersphere can be expressed as follows:
the constraint conditions are as follows: (h) i -m) T (h i -m)≤R 2 +ξ i ，i＝1，2，…，n
Using the lagrange multiplier method in combination with formula (11) and constraint conditions, we obtain:
lagrangian multiplier alpha i ＞0，γ i ＞0。
Solving the equation, for the above formula R, m, ζ i And respectively obtaining partial differentiation, enabling the partial differentiation to be equal to zero, and meeting the KKT condition as follows:
simplifying the Lagrangian multiplier to obtain:
L＝∑ i αi ( h i ·h i )-∑ i，j α i α j (h i ·h j ) (14)
if sample h i The following inequality is satisfied:
(h i -m) T (h i -m)≤R 2 +ξ i (15)
alpha is then i =0, indicating that the samples are inside the classification boundary.
If sample h i The following inequality is satisfied:
(h i -m) T (h i -m)＝R 2 +ξ i (16)
alpha is then i > 0, indicating that the sample is outside the classification boundary. The following results can be obtained:
the radius R of the hypersphere is the distance from the hypersphere center m to the boundary formed by any support vector, and can be determined by any support vector h sv And (3) obtaining:
when the support vector falls outside the described boundary, i.e. alpha i When=c, the sample is excluded due to the non-target sample.
When a test sample point q is newly added, the distance from the sample point q to the center m of the hypersphere can be defined as follows:
when R is q 2 ≤R 2 When the test sample belongs to the target sample, the model output value is 1; when R is q 2 ＞R 2 And when the model output value is 0, the test sample q is indicated to not belong to the target sample.
The output value of the SVDD model of the normal lifting appliance is 1, and the output value of the SVDD model of the fault lifting appliance is 0. Through the steps, the quick and accurate positioning and early warning of the fault lifting appliance can be realized according to the output value of the SVDD model.
2. The invention provides an intelligent fault diagnosis system of a production line lifting appliance based on PCA-SVDD, which comprises the following components: the device comprises a data transmission and storage module, a data preprocessing module, a data characteristic extraction module, a lifting appliance fault diagnosis module and a man-machine interaction module;
the data transmission and storage module is used for transmitting and storing the acquired digital, image and text data of the lifting appliance to a data processing center of the system;
the data preprocessing module processes the obtained data into a time window data form and preprocesses the problems of data deletion, noise, inconsistency and redundancy;
the data feature extraction module performs dimension reduction processing and feature extraction on the preprocessed data by using a PCA algorithm, and transmits the data to the lifting appliance fault diagnosis module;
the lifting appliance fault diagnosis module processes the obtained data by adopting an SVDD model and transmits the processing result to the man-machine interaction module;
the man-machine interaction module judges according to the processing result transmitted by the fault diagnosis module, accurately and rapidly judges the fault state of the lifting appliance and whether predictive maintenance is needed or not, and assists staff in making decisions.
The invention has the beneficial effects that:
1. the intelligent fault diagnosis method for the lifting appliance of the production line can realize quick and accurate positioning of the lifting appliance, greatly shortens the overhaul time, reduces the overhaul difficulty and has strong practicability. Compared with the conventional maintenance method, the method has the advantages that the shutdown loss is reduced greatly by almost millions of yuan per shutdown;
2. the sampling method based on the sliding time window adopted by the invention reduces the data volume required to be calculated by hundreds of times, so that the calculation process becomes simpler, and the calculation cost is also greatly reduced.
3. The sensitivity of the measurement is improved. The small faults can not cause the short circuit of the lifting appliance, but the accumulation of the small faults can cause the short circuit of the lifting appliance and the shutdown of the whole production line, the method and the system improve the detection capability of the small fault signals, and the maximum error time is not more than 0.4s.
4. The method reduces the false alarm rate and false alarm probability during fault detection, has higher prediction accuracy, and has positive significance for reducing the false alarm rate and false alarm rate of a detection system.
5. The robustness of the detection system is enhanced, the method does not involve the construction of a complex model, and the threshold value is not required to be established by means of expert experience, and each module of the system is relatively strong in mutual independence type but mutually unified, is convenient for unified management, and therefore has strong anti-interference capability.
6. The adaptation capability of the detection system is enhanced. The system can update the system by utilizing new information generated by the change, has self-learning capability, and can avoid mutual interference among the lifting appliances.
Drawings
FIG. 1 is a graph of voltage-current variation when a spreader fails;
FIG. 2 is a flow chart of a fault diagnosis method of the present invention;
FIG. 3 is a flow chart of sampling based on a sliding time window;
FIG. 4 is a diagram of the data output of the fault diagnosis result of the present invention (training set 100, test set 20);
FIG. 5 is a diagram of the data output of the fault diagnosis result of the present invention (training set 300, test set 200);
FIG. 6 is a diagram of the data output of the fault diagnosis result of the present invention (training set 500, test set 200);
FIG. 7 is a diagram of the data output of the fault diagnosis result of the present invention (training set 500, test set 500);
FIG. 8 is a diagram of the data output of the fault diagnosis result of the present invention (training set 700, test set 500);
FIG. 9 is a diagram of the data output of the fault diagnosis result of the present invention (training set 500, test set 700);
FIG. 10 is a diagram of the data output of the fault diagnosis result of the present invention (training set 1000, test set 700);
FIG. 11 is a flow chart of the operation of the fault diagnosis system of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to fig. 1 to 11 in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments herein without making any inventive effort are intended to fall within the scope of the present application.
Fig. 1 is a graph of voltage-current variation when a device fails. The image is obtained in a simulation mode, the time interval of simulation sampling is 0.002s, and the breaker is opened at the 2 nd s to protect the whole circuit under the assumption that the fault occurs at the 2 nd s. As can be seen from fig. 1, when the equipment has a short-circuit fault, the voltage and the current are obviously changed, and the equipment voltage data is selected to be adopted for fault diagnosis from the aspects of convenience in acquisition and diagnosis time.
The invention provides an intelligent fault diagnosis method for a production line lifting appliance based on PCA-SVDD. Fig. 2 is a flowchart of a specific embodiment of the intelligent fault diagnosis method of the production line lifting appliance based on PCA-SVDD. As can be seen from fig. 2, the intelligent fault diagnosis method for the production line lifting appliance based on PCA-SVDD of the present invention comprises the following specific steps:
data acquisition, transmission and storage
Sampling original voltage data of a lifting appliance by adopting industrial sensor equipment, wherein the acquisition period is T s =0.004 s, three-phase electricity of a, B, C is collected 1 time each in one period;
sampling based on sliding time window
The data is sampled using a sliding time window, and FIG. 3 is a flow chart based on sliding time window sampling. Through algorithm test, selecting the shape of the sliding time window as (100, 1), and the step length as 100;
data preprocessing
Preprocessing the acquired data, finding that the acquired voltage is in a (-5V, 5V) interval, mapping the voltage value to a (0, 65535) interval for convenience of acquisition, setting f (x) as a corresponding voltage value, and setting x as a specific numerical value, wherein the method comprises the following steps:
it should be noted that, the preprocessing mainly uses Python programming to process the problem of missing and repeated acquired data.
PCA-based data feature extraction
In order to realize fault diagnosis of the lifting appliance on the production line, a Principal Component Analysis (PCA) method is adopted to reduce the dimension of data, and when the data are processed, 300-dimension random vectors are firstly constructed, and X= (X) 1 ，X 2 ，……，X n ，……，X 300 ) (n is more than or equal to 1 and less than or equal to 300), sample data are selected, a sample matrix is constructed, and the following transformation is carried out: (1. Ltoreq.i, j. Ltoreq.n)
From the above variations, a normalized matrix a is obtained, and a correlation coefficient matrix R of the normalized matrix is obtained:
R＝[r ij ] p X p ＝A T A/(n-1) (4)
r ij ＝∑a ij ·a ij /(n-1) (5)
solving a characteristic root lambda of a characteristic equation of the matrix R, and determining a main component:
|R-λI p |＝0 (6)
for each lambda i J=1, 2, … …, m, solving the equation rb=λ j b, obtaining unitsFeature vectorConverting the standard variable into a main component:
wherein F is 1 Is the 1 st main component, F n Is the nth main component
SVDD-based equipment fault diagnosis
The first 17 main components are selected by using Python programming, and the extracted components are input into an SVDD model for fault diagnosis, and the operation result of the SVDD model is as follows:
the SVDD model can be summarized simply as:
the constraint conditions are: (h) i -A) T (h i -A)≤R 2 ，i＝1，2，…，n (9)
Wherein R represents the radius of the hypersphere, m represents the sphere center of the hypersphere, and n represents the number of sample data during training. Need to solve for the inclusion of all data h i Is a minimum hypersphere of (c).
The SVDD model finding the hypersphere can be summarized as a parameter iterative optimization problem (optimal solution problem), and the basic form is as follows:
in the constraint condition, a relaxation factor xi (non-negative parameter) and a penalty factor c are introduced.
After introducing the relaxation factor, the minimum problem of the hypersphere can be expressed as follows:
the constraint conditions are as follows: (h) i -m) T (h i -m)≤R 2 +ξ i ，i＝1，2，…，n
Using the lagrange multiplier method in combination with formula (11) and constraint conditions, we obtain:
lagrangian multiplier alpha i ＞0，γ i ＞0。
Solving the equation, for the above formula R, m, ζ i Partial differentiation (let partial differentiation equal zero) is separately calculated, and the KKT condition is satisfied:
simplifying the Lagrangian multiplier to obtain:
L＝∑ i α i (h i ·h i )-∑ i，j α i a j (h i ·h j ) (14)
if sample h i The following inequality is satisfied:
(h i -m) T (h i -m)≤R 2 +ξ i (15)
alpha is then i =0, indicating that the samples are inside the classification boundary.
If sample h i The following inequality is satisfied:
(h i -m) T (h i -m)＝R 2 +ξ i (16)
alpha is then i > 0, indicating that the sample is outside the classification boundary. The following results can be obtained:
the radius R of the hypersphere is the distance from the center m of the hypersphere to the boundary formed by the support vector, and can be communicatedPassing arbitrary support vector h sv And (3) obtaining:
when the support vector falls outside the described boundary, i.e. alpha i When=c, the sample is excluded due to the non-target sample.
When a test sample point q is newly added, the distance from the sample point q to the center m of the hypersphere can be defined as follows:
when R is q 2 ≤R 2 When the test sample belongs to the target sample, the model output value is 1; when R is q 2 ＞R 2 And when the model output value is 0, the test sample q is indicated to not belong to the target sample.
The output value of the SVDD model of the normal lifting appliance is 1, and the output value of the SVDD model of the fault lifting appliance is 0. Through the steps, the quick and accurate positioning and early warning of the fault lifting appliance can be realized according to the output value of the SVDD model.
In order to better illustrate the technical effects of the present invention, the effects of the present invention are experimentally verified using a specific example. Tens of millions of data of a lifting appliance of a production line of a final assembly workshop of a certain company in a normal state and a fault state are adopted in the embodiment.
In consideration of the actual requirements of feature extraction and fault diagnosis of the invention, five combined models of PCA-OCSVM, PCA-iForest, ICA-SVDD, ICA-OCSVM and UMAP-OCSVM are selected as comparison. In order to ensure the accuracy of the operation, seven test types of training 100 test groups 20, training 300 test groups 200, training 500 test groups 500, training 700 test groups 500, training 500 test groups 700 and training 1000 test groups 700 are selected. Fig. 4-10 are graphs of fault diagnosis result data output for seven test groups according to the present invention. The fault diagnosis of the lifting appliance on the production line of the assembly workshop is carried out by adopting the method and other methods respectively, and the results are shown in the following table 1:
table 1: precision comparison table for fault diagnosis by different methods
As can be seen from Table 1, the fault diagnosis accuracy of the invention is higher than that of other algorithms, the highest diagnosis accuracy is maintained to be more than 94%, and the highest diagnosis accuracy is 99.3%, thus the invention has excellent effect in fault diagnosis of the lifting appliance on the automobile production line.
The invention provides an intelligent fault diagnosis system of a lifting appliance of a production line based on PCA-SVDD, which comprises a data transmission and storage module, a data preprocessing module, a data feature extraction module, a lifting appliance fault diagnosis module and a man-machine interaction module. FIG. 4 is a flow chart of the operation of the fault diagnosis system of the present invention.
1. The data transmission and storage module is responsible for transmitting and storing the acquired data to a data processing center of the system;
2. the data preprocessing module is used for processing the obtained data into a time window data form and preprocessing the problems of data loss, noise, inconsistency, redundancy and the like;
3. the data feature extraction module performs dimension reduction processing and feature extraction on the preprocessed data by using a PCA algorithm, and transmits the data to the lifting appliance fault diagnosis module;
4. the lifting appliance fault diagnosis module processes the obtained data by adopting an SVDD model and transmits the processing result to the man-machine interaction module;
5. the man-machine interaction module judges according to the processing result transmitted by the fault diagnosis module, accurately and rapidly judges the fault state of the lifting appliance and whether predictive maintenance is needed or not, and assists staff in making decisions.
While the foregoing describes the embodiments of the present invention so as to facilitate the understanding of the present invention by those skilled in the art, it should be apparent that the present invention is not limited to the scope of the embodiments, but is to be construed as being protected by the accompanying claims insofar as various changes are within the spirit and scope of the present invention as defined and defined by the appended claims.
Claims (2)
1. An intelligent fault diagnosis method for a production line lifting appliance based on PCA-SVDD is characterized by comprising the following steps:
step 1: sampling original voltage data of the equipment, wherein the acquisition period is T s ∈[0.0004s，1s]Three-phase electricity of A, B and C is collected 1 time in a period;
step 2: sampling data by adopting a sliding time window, and selecting the shape of the sliding time window as (100, 1) and the step length as 100 through algorithm test;
step 3: preprocessing the acquired data, finding that the acquired voltage is in a (-5V, 5V) interval, mapping the voltage value to a (0, 65535) interval for convenience of acquisition, setting f (x) as a corresponding voltage value, and setting x as a specific numerical value, wherein the method comprises the following steps:
step 4: the method comprises the steps of performing dimensionality reduction on data by adopting a Principal Component Analysis (PCA), and firstly constructing 300-dimensional random vectors and X= (X) when processing 1 ，X 2 ，……，X n ，……，X 300 ) (n is more than or equal to 1 and less than or equal to 300), sample data are selected, a sample matrix is constructed, and the following transformation is carried out: (1. Ltoreq.i, j. Ltoreq.n)
From the above variations, a normalized matrix a is obtained, and a correlation coefficient matrix R of the normalized matrix is obtained:
R＝[r ij ] p X p ＝A T A/(n-1) (4)
r ij ＝∑a ij ·a ij /(n-1) (5)
solving a characteristic root lambda of a characteristic equation of the matrix R, and determining a main component:
|R-λI p |＝0 (6)
for each lambda j J=1, 2,, m, solving the equation rb=λ j b, obtaining unit feature vectorConverting the standard variable into a main component:
formula (8) wherein F 1 Is the 1 st main component, F n Is the nth principal component and is used as the main component,
step 5: the first 17 main components are selected by using Python programming, and the extracted components are input into an SVDD model for fault diagnosis, and the operation result of the SVDD model is as follows:
the SVDD model can be summarized simply as:
the constraint conditions are: (h) i -A) T (h i -A)≤R 2 ,i＝1，2，…，n (9)
In the formula (9), R represents the radius of the hypersphere, m represents the sphere center of the hypersphere, and n represents the number of sample data during training. Need to solve for the inclusion of all data h i Is a minimum hypersphere of (c).
The SVDD model searching hypersphere can be summarized as a parameter iteration optimization problem, and the basic form is as follows:
and in the constraint condition, introducing a non-negative parameter relaxation factor xi and a penalty factor c.
After introducing the relaxation factor, the minimum problem of the hypersphere can be expressed as follows:
the constraint conditions are as follows: (h) i -m) T (h i -m)≤R 2 +ξ i ，i＝1，2，…，n
Using the lagrange multiplier method in combination with formula (11) and constraint conditions, we obtain:
lagrangian multiplier alpha i ＞0，γ i ＞0。
Solving the equation, for the above formula R, m, ζ i And respectively obtaining partial differentiation, enabling the partial differentiation to be equal to zero, and meeting the KKT condition as follows:
simplifying the Lagrangian multiplier to obtain:
L＝∑ i a i (h i ·h i )-∑ i，j α i α j (h i ·h j ) (14)
if sample h i The following inequality is satisfied:
(h i -m) T (h i -m)≤R 2 +ξ i (15)
alpha is then i =0, indicating that the samples are inside the classification boundary.
If sample h i The following inequality is satisfied:
(h i -m) T (h i -m)＝R 2 +ξ i (16)
alpha is then i > 0, indicating that the sample is outside the classification boundary. The following results can be obtained:
the radius R of the hypersphere is the distance from the hypersphere center m to the boundary formed by any support vector, and can be determined by any support vector h sv And (3) obtaining:
when the support vector falls outside the described boundary, i.e. alpha i When=c, the sample is excluded due to the non-target sample.
When a test sample point q is newly added, the distance from the sample point q to the center m of the hypersphere can be defined as follows:
when R is q 2 ≤R 2 When the test sample belongs to the target sample, the model output value is 1; when R is q 2 ＞R 2 And when the model output value is 0, the test sample q is indicated to not belong to the target sample.
The output value of the SVDD model of the normal lifting appliance is 1, and the output value of the SVDD model of the fault lifting appliance is 0. Through the steps, the quick and accurate positioning and early warning of the fault lifting appliance can be realized according to the output value of the SVDD model.
2. An intelligent fault diagnosis system of a production line lifting appliance based on PCA-SVDD is characterized by comprising the following components: the device comprises a data transmission and storage module, a data preprocessing module, a data characteristic extraction module, a lifting appliance fault diagnosis module and a man-machine interaction module;
the data transmission and storage module is used for transmitting and storing the acquired digital, image and text data of the lifting appliance to a data processing center of the system;
the data preprocessing module processes the obtained data into a time window data form and preprocesses the problems of data deletion, noise, inconsistency and redundancy;
the data feature extraction module performs dimension reduction processing and feature extraction on the preprocessed data by using a PCA algorithm, and transmits the data to the lifting appliance fault diagnosis module;
the lifting appliance fault diagnosis module processes the obtained data by adopting an SVDD model and transmits the processing result to the man-machine interaction module;
the man-machine interaction module judges according to the processing result transmitted by the fault diagnosis module, accurately and rapidly judges the fault state of the lifting appliance and whether predictive maintenance is needed or not, and assists staff in making decisions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310259582.4A CN117407757A (en) | 2023-03-16 | 2023-03-16 | Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310259582.4A CN117407757A (en) | 2023-03-16 | 2023-03-16 | Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117407757A true CN117407757A (en) | 2024-01-16 |
Family
ID=89491416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310259582.4A Pending CN117407757A (en) | 2023-03-16 | 2023-03-16 | Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117407757A (en) |
-
2023
- 2023-03-16 CN CN202310259582.4A patent/CN117407757A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113156917B (en) | Power grid equipment fault diagnosis method and system based on artificial intelligence | |
| CN110441065B (en) | Gas turbine on-line detection method and device based on LSTM | |
| CN112884089A (en) | Power transformer fault early warning system based on data mining | |
| CN114676742A (en) | Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network | |
| CN116614177B (en) | Optical fiber state multidimensional parameter monitoring system | |
| CN112906764A (en) | Communication safety equipment intelligent diagnosis method and system based on improved BP neural network | |
| CN111273125A (en) | RST-CNN-based power cable channel fault diagnosis method | |
| CN115392782A (en) | Method and system for monitoring and diagnosing health state of process system of nuclear power plant | |
| CN110850837A (en) | System life analysis and fault diagnosis method based on long-time and short-time memory neural network | |
| CN117371207A (en) | Extra-high voltage converter valve state evaluation method, medium and system | |
| CN117032165A (en) | Industrial equipment fault diagnosis method | |
| CN115564075B (en) | Main and auxiliary integrated fault collaborative diagnosis method and system for urban power grid | |
| CN117407757A (en) | Intelligent fault diagnosis method and system for production line lifting appliance based on PCA-SVDD | |
| CN115828744A (en) | White light LED fault on-line diagnosis and service life prediction method | |
| CN114626482A (en) | Active digital array radar real-time fault diagnosis method based on deep learning | |
| CN114137915A (en) | Fault diagnosis method for industrial equipment | |
| CN114487643A (en) | On-spot handing-over of extra-high voltage GIL equipment is accepted and is synthesized test platform | |
| CN112418398A (en) | Safety monitoring method for power information equipment | |
| CN113076217A (en) | Disk fault prediction method based on domestic platform | |
| CN115556099B (en) | Sustainable learning industrial robot fault diagnosis system and method | |
| Zheng et al. | Research on Predicting Remaining Useful Life of Equipment Based on Health Index | |
| CN117688496B (en) | Abnormality diagnosis method, system and equipment for satellite telemetry multidimensional time sequence data | |
| CN114755920B (en) | Environmental safety monitoring system for biosafety laboratory | |
| CN116028849B (en) | Emulsion pump fault diagnosis method based on depth self-coding network | |
| CN117370870B (en) | Knowledge and data compound driven equipment multi-working condition identification and performance prediction method |
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 |