CN113379252A - Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy - Google Patents
Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy Download PDFInfo
- Publication number
- CN113379252A CN113379252A CN202110654420.1A CN202110654420A CN113379252A CN 113379252 A CN113379252 A CN 113379252A CN 202110654420 A CN202110654420 A CN 202110654420A CN 113379252 A CN113379252 A CN 113379252A
- Authority
- CN
- China
- Prior art keywords
- switch cabinet
- voltage switch
- scale
- value
- arrangement entropy
- 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
- 230000036541 health Effects 0.000 title claims abstract description 44
- 238000011156 evaluation Methods 0.000 title claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 4
- 238000013210 evaluation model Methods 0.000 claims description 23
- 230000008569 process Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000008030 elimination Effects 0.000 claims description 6
- 238000003379 elimination reaction Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims 1
- 239000004576 sand Substances 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000005070 sampling Methods 0.000 abstract description 3
- 238000000605 extraction Methods 0.000 abstract 1
- 238000012423 maintenance Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013398 bayesian method Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 239000012774 insulation material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
A health system assessment method of a Bayesian network high-voltage switch cabinet based on multi-scale arrangement entropy comprises the steps of firstly, analyzing and fusing information of a plurality of physical sensors of the high-voltage switch cabinet; secondly, effective extraction of characteristic information is realized by introducing coarse graining operation and an arrangement entropy method; and finally, constructing and training a Bayesian network model based on expert domain knowledge to realize the health state evaluation of the high-voltage switch cabinet system. According to the invention, data obtained by sampling by different detection means can be fused, so that the safety state of the high-voltage switch cabinet can be accurately evaluated, and the safe and stable operation of the electric power high-voltage switch cabinet can be ensured.
Description
Technical Field
The invention belongs to the field of state evaluation of high-voltage switch cabinets, and particularly relates to a method for evaluating a health system of an electric power high-voltage switch cabinet, which is suitable for ensuring the normal operation of the electric power high-voltage switch cabinet system.
Background
The operation state of the power high-voltage switch cabinet, which is one of the main devices of the power system, is related to the safety of the whole power system. The condition maintenance is a necessary means for ensuring the safe and reliable operation of the electric power high-voltage switch cabinet, and the core and the foundation of the condition maintenance are condition evaluation. Through carrying out the comprehensive assessment to electric power high tension switchgear, and then formulate the maintenance strategy according to the assessment result, not only can improve electric power high tension switchgear's operational reliability, also can produce huge economic benefits.
In the operation of the electric high-voltage switch cabinet, partial discharge and even combustion and explosion of the electric high-voltage switch cabinet can be caused due to the reasons of moisture, insufficient insulation distance, deterioration of an insulation material and the like. At present, means for detecting the electrification of the power high-voltage switch cabinet mainly comprise an infrared diagnosis method, a transient earth voltage method, an ultrasonic method, an ultrahigh frequency method and the like. These methods generally determine the safety status level of a high-voltage switchgear by detecting the sound, light and electrical signals generated when a power high-voltage switchgear fails. The above method is widely used in industrial fields, but is very vulnerable to field electromagnetic and mechanical interference. In addition, the above methods are too dependent on subjective judgment and have limitations, and if only one of the methods is relied on, misjudgment is often caused.
The Bayesian network provides a new approach for the health system evaluation of the power high-voltage switch cabinet, and the Bayesian method is widely applied to the system fault diagnosis and evaluation. Therefore, the invention integrates the relevant live detection data, environmental factors and equipment self factors of the high-voltage power switch cabinet, and provides a Bayesian network model based on multi-scale permutation entropy for health system evaluation of the high-voltage power switch cabinet. Through embedding expert knowledge and analyzing historical data obtained by detecting the power high-voltage switch cabinet, a Bayesian network model is learned, constructed and established from the data, and the established Bayesian network model is utilized to realize the safety state evaluation of the power high-voltage switch cabinet so as to guide maintenance.
Disclosure of Invention
Aiming at the technical defects of the existing health state evaluation method of the electric power high-voltage switch cabinet system, the health state evaluation method of the Bayesian network electric power high-voltage switch cabinet based on the multi-scale arrangement entropy is provided. According to the invention, data obtained by sampling by different detection means can be fused, so that the safety state of the high-voltage switch cabinet can be accurately evaluated, and the safe and stable operation of the electric power high-voltage switch cabinet can be ensured.
The technical scheme adopted by the invention is as follows:
a health system assessment method for a Bayesian network power high-voltage switch cabinet based on multi-scale arrangement entropy values comprises the following steps:
1) the model training stage comprises the following steps:
1.1) respectively acquiring working time sequence data of a historical normal power high-voltage switch cabinet under h sensors, wherein the normal working condition time sequence data acquired by a single sensor under an offline condition is { x }h(i),i＝1,2,…,n；h∈N+And coarsely granulating the time sequence data, namely:
wherein the content of the first and second substances,
obtaining a multi-scale time sequence with the time of w after coarse graining treatment, wherein N is the length of the time sequence, s is a scale factor, and s belongs to N+，N+The characteristic information under different scale factors s is obtained through coarse graining treatment to improve the evaluation performance of the model;
1.2) calculating a multi-scale time series arrangement entropy value under a single sensor;
1.3) similarly, respectively calculating the arrangement entropy values of the multi-scale time sequence under h sensors according to the step 1.2), and carrying out equal-interval discretization on the obtained multi-scale arrangement entropy value data, namely, firstly, determining the range [ a, b ] of the multi-scale time sequence arrangement entropy values by searching the minimum value a and the maximum value b in the data]Next, M (M.gtoreq.2 and M ∈ N) is performed for this range+) Reasonably divide the equal interval into intervals with the length of
Finally, each subinterval is assigned with a label value of 0,1,2 and …, and the multi-scale permutation entropy value set obtained by definition is { eh,s，h,s＝1,2,3,…}；
1.4) constructing a health system evaluation model structure of the high-voltage power switch cabinet according to bus chamber state information, cable chamber state information, vacuum load switch state information and fusion expert prior knowledge in the high-voltage power switch cabinet, namely determining topological relations among node variables in the health system evaluation model of the high-voltage power switch cabinet to form a directed acyclic graph;
1.5) based on the structure of the health system evaluation model of the high-voltage power switch cabinet constructed in the step 1.4), determining parameters among relative node variables in the evaluation model structure by Bayesian estimation, namely a conditional probability table among the node variables, firstly, taking the parameters to be estimated as a random variable, allowing the parameters to obey the probability distribution of Dirichlet, then, learning according to the prior distribution of the parameters and a multi-scale arrangement entropy data set of time series acquired by different sensors in the high-voltage power switch cabinet, solving the posterior distribution of the parameters, and finally, solving the expected values of the parameters as the final values of the parameters. Assuming that a sample set D of Q sequences is obtained, an observation value is defined as U ═ U (U)1,U2,U3,…,UQ)；
2) The threshold determination phase is divided into three steps:
2.1) respectively calculating the multi-scale arrangement entropy values of the normal working condition time sequence under h sensors in the electric high-voltage switch cabinet, and performing equal-interval discretization operation on the multi-scale arrangement entropy values according to the step 1.3) to obtain a multi-scale arrangement entropy set { e }h,s，h,s＝1,2,3,…}；
2.2) processing the multiscale arrangement entropy value data set { e) obtained in the step 2.1)h,sH, s is 1,2,3, …, inputting to the health system evaluation model of the high-voltage power switch cabinet constructed in the step 1);
2.3) determining the threshold by MarkovThe variable elimination inference engine of the network deduces to obtain the posterior probability value P (T) corresponding to each momentq|E＝{eh,s}), namely:
expression (2) indicates that the feature information set { e) is inputh,sAfter h, s ═ 1,2,3, … } the target is Tq(q ═ 1 or-1), where P (E ═ E)h,s}) is the prior probability, Tq(q-1 or-1) corresponds to the state of the power high-voltage switchgear system (q-1 represents healthy, q-1 represents abnormal). By taking the logarithm of equation (2) and then selecting the minimum value as the threshold R:
R＝min(-ln(P(Tq|E＝{eh,s}))) (3)
3) the health evaluation stage comprises the following steps:
3.1) calculating the multi-scale arrangement entropy value of the time sequence acquired under h sensors in the electric high-voltage switch cabinet in real time, and similarly, performing equal-interval discretization operation on the multi-scale arrangement entropy value according to the step 1.3) to obtain a multi-scale arrangement entropy data set { eh,s，h,s＝1,2,3,…}；
3.2) processing the multiscale permutation entropy data set { e) obtained in 3.1) in real timeh,sInputting h, s-1, 2,3, … } into the evaluation model of the health system of the high-voltage power switch cabinet constructed in the step 1), selecting a Markov network variable elimination inference engine to calculate a probability value P (T) corresponding to each momentq|E＝{eh,s}), namely:
3.3) real-time health assessment.
Further, in said 3.3), P (T) obtained in the above stepq|E＝{eh,sGet the logarithm value of the power high-voltage switch cabinet, compare the logarithm value with a threshold value, and if the logarithm value exceeds the threshold value R, the health state of the power high-voltage switch cabinet is considered to be in a problem, so that the power high-voltage switch cabinet needs to be timely judgedOverhauling; and if the threshold value R is not exceeded, the electric high-voltage switch cabinet is considered to be operated normally.
Still further, the process of 1.2) is as follows:
1.2.1) multiscale time series
Performing phase space reconstruction, namely:
in the formula (I), the compound is shown in the specification,
the time sequence is the time sequence with the moment j after the original multi-scale time sequence is subjected to phase space reconstruction, wherein mu is the embedding dimension, and lambda is the delay time.
1.2.2) when the embedding dimension μ is real, all reconstruction component elements in step 1.2.1) are re-ordered according to the magnitude as follows:
1.2.3) counting the probability of occurrence of K symbol sequences S (g), and marking the symbol sequences as { P }gAnd g is 1,2,3, …, K, calculating the Shannon entropy and carrying out normalization processing, and defining the arrangement entropy e of the coarse-grained time series under different sensors in the high-voltage switch cabinet as,
in the process of calculating the multi-scale permutation entropy, a scale factor s is flexibly selected according to the required characteristic dimension, and an embedding dimension mu and a delay time lambda need to be determined before permutation entropy calculation is carried out.
Further, the process of 1.5) is as follows:
1.5.1) first a prior distribution P (θ) of the network parameter θ is determined, where P (θ) obeys an m-dimensional dirichlet probability distribution, i.e.:
wherein Dir is Dirichlet distribution, m is Dirichlet distribution dimension, m belongs to [1, r ], alpha is hyper parameter, and tau (alpha) is Gama function;
1.5.2) the probability of a sample occurring is:
1.5.3) solving the posterior distribution of theta by using a Bayesian formula:
wherein Q ismThe evaluation model parameters are Q sequences contained in the m-dimensional Dirichlet distribution sample set, the determined network parameters theta are the high-voltage switch cabinet health system evaluation model parameters with the maximum posterior probability, and the expected values of the parameters are calculated to serve as final values.
The method of the invention has the following advantages: 1. the information of a plurality of physical sensors in the electric power high-voltage switch cabinet system is analyzed, so that the effect of the method is better than that of a method based on a single sensor. 2. Through coarse graining operation, more characteristic information can be obtained with little calculation cost, and certain detection performance is improved while the real-time performance of the electric power high-voltage switch cabinet system is met. 3. The invention combines the prior knowledge in the expert field, so that the model has better generalization capability.
Drawings
FIG. 1 is a flow chart of a multi-scale arrangement entropy calculation of a time series acquired by an electric high-voltage switch cabinet;
FIG. 2 is a diagram of a Bayesian network of a health assessment system for a power high-voltage switch cabinet;
FIG. 3 is a flow chart of Bayesian network accurate inference of a health assessment system of a power high-voltage switch cabinet;
FIG. 4 is a flow chart of a method for evaluating the state of a power high-voltage switch cabinet;
fig. 5 is a schematic diagram of a health status evaluation method of the power high-voltage switch cabinet.
Detailed Description
In order to make the technical scheme and the design idea of the present invention clearer, we describe the present invention in detail with reference to the accompanying drawings.
The research object is an electric power high-voltage switch cabinet, and the sampling data is parameter information acquired by a sensor in the electric power high-voltage switch cabinet.
Referring to fig. 1, a multi-scale arrangement entropy calculation process of a working condition time sequence of an electric high-voltage switch cabinet generally includes performing coarse graining processing on a time sequence acquired by each sensor in the electric high-voltage switch cabinet, then performing sequence reconstruction on the processed coarse graining sequence, and finally calculating an arrangement entropy value of the sequence according to a calculation method of the arrangement entropy, namely a multi-scale arrangement entropy value.
Referring to fig. 4, a method for evaluating a health system of a high-voltage power switch cabinet based on a multi-scale arrangement entropy and a bayesian network comprises the following steps:
1) the model training stage comprises the following steps:
1.1) respectively acquiring working time sequence data of a historical normal power high-voltage switch cabinet under h sensors, wherein the normal working condition time sequence data acquired by a single sensor under an offline condition is { x }h(i),i＝1,2,…,n；h∈N+And coarsely granulating the time sequence data, namely:
wherein the content of the first and second substances,
a multi-scale time sequence with the time of w is obtained after coarse graining treatment, n is the length of the time sequence, s is a scale factor, ands∈N+，N+the characteristic information under different scale factors s is obtained through coarse graining treatment to improve the evaluation performance of the model;
1.2) calculating the multi-scale time series arrangement entropy under a single sensor by the following process:
1.2.1) multiscale time series
Performing phase space reconstruction, namely:
in the formula (I), the compound is shown in the specification,
the time sequence is a time sequence with the moment j after the original multi-scale time sequence is subjected to phase space reconstruction, wherein mu is an embedding dimension, and lambda is delay time;
1.2.2) when the embedding dimension μ is real, all reconstruction component elements in step 1.2.1) are re-ordered according to the magnitude as follows:
1.2.3) counting the probability of occurrence of K symbol sequences S (g), and marking the symbol sequences as { P }gAnd g is 1,2,3, …, K, calculating the Shannon entropy and carrying out normalization processing, and defining the arrangement entropy e of the coarse grained time series under different sensors in the electric high-voltage switch cabinet as,
in the process of calculating the multi-scale arrangement entropy value, a scale factor s is flexibly selected according to required characteristic dimensions, an embedding dimension mu and a delay time lambda are required to be determined before arrangement entropy calculation is carried out, in the invention, mu belongs to [3,7], and lambda is 1;
1.3) similarly, respectively calculating the arrangement entropy values of the multi-scale time sequence under h sensors according to the step 1.2), and performing equal-interval discretization on the obtained multi-scale arrangement entropy data, namely, firstly determining the range [ a, b ] of the arrangement entropy values of the multi-scale time sequence by searching the minimum value a and the maximum value b in the data]Next, M (M.gtoreq.2 and M ∈ N) is performed for this range+) Reasonably divide the equal interval into intervals with the length of
Finally, each subinterval is assigned with a label value of 0,1,2 and …, and the multi-scale permutation entropy value set obtained by definition is { eh,s，h,s＝1,2,3,…}；
1.4) with reference to fig. 2, according to bus chamber state information, cable chamber state information and vacuum load switch state information in the electric high-voltage switch cabinet and fusion expert prior knowledge, constructing an electric high-voltage switch cabinet health system evaluation model structure, namely determining topological relation among node variables in the electric high-voltage switch cabinet health system evaluation model to form a directed acyclic graph;
1.5) referring to fig. 5, based on the structure of the evaluation model of the health system of the high-voltage power switch cabinet constructed in the step 1.4), determining parameters among relative node variables in the evaluation model structure by using bayesian estimation, namely, a conditional probability table among the node variables, firstly, using the parameters to be estimated as a random variable, allowing the parameters to obey dirichlet probability distribution, then, learning according to prior distribution of the parameters and a multi-scale arrangement entropy data set of time sequences acquired by different sensors in the high-voltage power switch cabinet, calculating posterior distribution of the parameters, finally, calculating expected values of the parameters as final values, assuming that sample sets D of Q sequences are acquired, and defining an observation value as U-U (U-U)1,U2,U3,…,UQ) (ii) a The process is as follows:
1.5.1) first a prior distribution P (θ) of the network parameter θ is determined, where P (θ) obeys an m-dimensional dirichlet probability distribution, i.e.:
wherein Dir is Dirichlet distribution, m is Dirichlet distribution dimension, m belongs to [1, r ], alpha is hyper parameter, and tau (alpha) is Gama function;
1.5.2) the probability of a sample occurring is:
1.5.3) solving the posterior distribution of theta by using a Bayesian formula:
wherein Q ismThe method comprises the steps that Q sequences contained in an m-dimensional Dirichlet distribution sample set are obtained, the determined network parameter theta is the parameter of the health system evaluation model of the high-voltage power switch cabinet when the posterior probability is maximum, and the expected value of the parameter is calculated to serve as the final value;
2) referring to fig. 4 and 5, the threshold determination phase is divided into the following steps:
2.1) respectively calculating the multi-scale arrangement entropy values of the normal working condition time sequence under h sensors in the electric high-voltage switch cabinet, and performing equal-interval discretization operation on the multi-scale arrangement entropy values according to the step 1.3) to obtain a multi-scale arrangement entropy set { e }h,s，h,s＝1,2,3,…}；
2.2) processing the multiscale arrangement entropy value data set { e) obtained in the step 2.1)h,sH, s is 1,2,3, …, inputting to the health system evaluation model of the high-voltage power switch cabinet constructed in the step 1);
2.3) determining the threshold, and referring to FIG. 3, deducing the posterior probability value P (T) corresponding to each moment by a variable elimination inference engine of the Markov networkq|E＝{eh,s}), namely:
expression (2) indicates that the feature information set { e) is inputh,sAfter h, s ═ 1,2,3, … } the target is Tq(q ═ 1 or-1), where P (E ═ E)h,s}) is the prior probability, Tq(q-1 or-1) corresponds to the state of the power high-voltage switchgear system (q-1 represents healthy, q-1 represents abnormal). By taking the logarithm of equation (2) and then selecting the minimum value as the threshold R:
R＝min(-ln(P(Tq|E＝{eh,s}))) (9)
3) the health evaluation stage comprises the following steps:
3.1) with reference to fig. 1 and 4, calculating the multi-scale arrangement entropy values of the time sequence acquired under h sensors in the electric high-voltage switch cabinet in real time, and similarly, performing equal-interval discretization operation on the multi-scale arrangement entropy values according to the step 1.3) to obtain a multi-scale arrangement entropy data set { eh,s，h,s＝1,2,3,…}；
3.2) referring to FIG. 5, the multiscale permuted entropy data set { e } processed in 3.1) is processed in real timeh,sInputting h, s-1, 2,3, … } into the evaluation model of the health system of the high-voltage power switch cabinet constructed in the step 1), selecting a Markov network variable elimination inference engine to calculate a probability value P (T) corresponding to each momentq|E＝{eh,s}), namely:
3.3) real-time health assessment.
Further, in said 3.3), P (T) obtained in the above stepq|E＝{eh,sThe logarithm value is taken, the logarithm value is compared with a threshold value, if the logarithm value exceeds the threshold value R, the health state of the electric power high-voltage switch cabinet is considered to have a problem, and the electric power high-voltage switch cabinet needs to be overhauled in time; and if the threshold value R is not exceeded, the electric high-voltage switch cabinet is considered to be operated normally.
The embodiments described in this specification are merely illustrative of implementations of the inventive concepts, which are intended for purposes of illustration only. The scope of the present invention should not be construed as being limited to the particular forms set forth in the examples, but rather as being defined by the claims and the equivalents thereof which can occur to those skilled in the art upon consideration of the present inventive concept.
Claims (4)
1. A health system assessment method of a Bayesian network high-voltage switch cabinet based on multi-scale arrangement entropy is characterized by comprising the following steps:
1) the model training stage comprises the following steps:
1.1) respectively acquiring historical normal high-voltage switch cabinet working time sequence data under h sensors, wherein the normal working condition time sequence data acquired by a single sensor under the offline condition is { x }h(i),i＝1,2,…,n；h∈N+And coarsely granulating the time sequence data, namely:
wherein the content of the first and second substances,
obtaining a multi-scale time sequence with the time of w after coarse graining treatment, wherein N is the length of the time sequence, s is a scale factor, and s belongs to N+，N+The characteristic information under different scale factors s is obtained through coarse graining treatment to improve the evaluation performance of the model;
1.2) calculating a multi-scale time series arrangement entropy value under a single sensor;
1.3) similarly, respectively calculating the arrangement entropy values of the multi-scale time sequence under h sensors according to the step 1.2), and performing equal-interval discretization on the obtained multi-scale arrangement entropy data, namely, firstly determining the range [ a, b ] of the arrangement entropy values of the multi-scale time sequence by searching the minimum value a and the maximum value b in the data]Next, M (M.gtoreq.2 and M ∈ N) is performed for this range+) Reasonably divide the equal interval into intervals with the length of
Finally, each subinterval is assigned with a label value of 0,1,2 and …, and the multi-scale permutation entropy value set obtained by definition is { eh,s，h,s＝1,2,3,…}；
1.4) constructing a health system evaluation model structure of the high-voltage switch cabinet according to bus chamber state information, cable chamber state information, vacuum load switch state information and fusion expert prior knowledge in the high-voltage switch cabinet, namely determining a topological relation among node variables in the health system evaluation model of the high-voltage switch cabinet to form a directed acyclic graph;
1.5) based on the structure of the high-voltage switch cabinet health system evaluation model constructed in the step 1.4), determining parameters among relative node variables in the evaluation model structure by using Bayesian estimation, namely a conditional probability table among the node variables, firstly, taking the parameters to be estimated as a random variable, allowing the parameters to obey the probability distribution of Dirichlet, then, learning according to the prior distribution of the parameters and a multi-scale arrangement entropy data set of time sequences acquired by different sensors in the high-voltage switch cabinet, calculating the posterior distribution of the parameters, finally, calculating expected values of the parameters as final values, assuming that a sample set D of Q sequences is acquired, and defining an observed value as U ═ (U ═ by calculating the expected values of the parameters as the final values of the sample set D1,U2,U3,…,UQ)；
2) A threshold determination phase comprising the steps of:
2.1) respectively calculating the multi-scale arrangement entropy values of the normal working condition time sequence under h sensors in the high-voltage switch cabinet, and performing equal-interval discretization operation on the multi-scale arrangement entropy values according to the step 1.3) to obtain a multi-scale arrangement entropy set { eh,s，h,s＝1,2,3,…}；
2.2) processing the multiscale arrangement entropy value data set { e) obtained in the step 2.1)h,sH, s is 1,2,3, …, inputting the h, s is into the high-voltage switch cabinet health system evaluation model constructed in the step 1);
2.3) determining a threshold value, and obtaining a posterior probability value P (T) corresponding to each moment by reasoning through a variable elimination reasoning engine of the Markov networkq|E＝{eh,s}), namely:
expression (2) indicates that the feature information set { e) is inputh,sAfter h, s ═ 1,2,3, … } the target is TqWhere P (E ═ { E ═ E)h,s}) is the prior probability, TqCorresponding to the state of the high-voltage switch cabinet system, q is 1 for health, q is-1 for abnormality, and the minimum value is selected as the threshold value R by taking the logarithm value of the formula (2):
R＝min(-ln(P(Tq|E＝{eh,s}))) (9)
3) the health evaluation stage comprises the following steps:
3.1) calculating the multi-scale arrangement entropy value of the time sequence acquired under h sensors in the high-voltage switch cabinet in real time, and similarly, performing equal-interval discretization operation on the multi-scale arrangement entropy value according to the step 1.3) to obtain a multi-scale arrangement entropy data set { eh,s，h,s＝1,2,3,…}；
3.2) processing the multiscale permutation entropy data set { e) obtained in 3.1) in real timeh,sH, s is 1,2,3, …, inputting to the evaluation model of the high-voltage switch cabinet health system constructed in the step 1), selecting a Markov network variable elimination inference engine to calculate a probability value P (T) corresponding to each momentq|E＝{eh,s}), namely:
3.3) real-time health assessment.
2. The method for evaluating the health system of the Bayesian network high-voltage switch cabinet based on multi-scale arrangement entropy of claim 1, wherein in the step 3.3), P (T) is obtained according to the previous stepq|E＝{eh,sAnd) taking a logarithmic value, comparing the logarithmic value with a threshold value, and if the logarithmic value exceeds the threshold value, judging that the operation of the high-voltage switch cabinet is abnormal.
3. The method for evaluating the health system of the Bayesian network high-voltage switch cabinet based on the multi-scale arrangement entropy as set forth in claim 1 or 2, wherein the process of 1.2) is as follows:
1.2.1) multiscale time series
Performing phase space reconstruction, namely:
in the formula (I), the compound is shown in the specification,
the time sequence is the time sequence with the moment j after the original multi-scale time sequence is subjected to phase space reconstruction, wherein mu is the embedding dimension, and lambda is the delay time.
1.2.2) when the embedding dimension μ is real, all reconstruction component elements in step 1.2.1) are re-ordered according to the magnitude as follows:
1.2.3) counting the probability of occurrence of K symbol sequences S (g), and marking the symbol sequences as { P }gAnd g is 1,2,3, …, K, calculating the Shannon entropy and carrying out normalization processing, and defining the arrangement entropy e of the coarse-grained time series under different sensors in the high-voltage switch cabinet as,
in the process of calculating the multi-scale permutation entropy, a scale factor s is flexibly selected according to the required characteristic dimension, and an embedding dimension mu and a delay time lambda need to be determined before permutation entropy calculation is carried out.
4. The method for evaluating the health system of the Bayesian network high-voltage switch cabinet based on the multi-scale arrangement entropy as set forth in claim 1 or 2, wherein the process of 1.5) is as follows:
1.5.1) first a prior distribution P (θ) of the network parameter θ is determined, where P (θ) obeys an m-dimensional dirichlet probability distribution, i.e.:
wherein Dir is Dirichlet distribution, m is Dirichlet distribution dimension, m belongs to [1, r ], alpha is hyper parameter, and tau (alpha) is Gama function;
1.5.2) the probability of a sample occurring is:
1.5.3) solving the posterior distribution of theta by using a Bayesian formula:
wherein Q ismThe evaluation model parameters are Q sequences contained in the m-dimensional Dirichlet distribution sample set, the determined network parameters theta are the high-voltage switch cabinet health system evaluation model parameters with the maximum posterior probability, and the expected values of the parameters are calculated to serve as final values.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110654420.1A CN113379252A (en) | 2021-06-11 | 2021-06-11 | Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110654420.1A CN113379252A (en) | 2021-06-11 | 2021-06-11 | Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113379252A true CN113379252A (en) | 2021-09-10 |
Family
ID=77574003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110654420.1A Pending CN113379252A (en) | 2021-06-11 | 2021-06-11 | Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113379252A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113554131A (en) * | 2021-09-22 | 2021-10-26 | 四川大学华西医院 | Medical image processing and analyzing method, computer device, system and storage medium |
| CN114167838A (en) * | 2021-12-03 | 2022-03-11 | 西安电子科技大学 | Multi-scale health assessment and fault prediction method for servo system |
| CN117056086A (en) * | 2023-10-11 | 2023-11-14 | 国网山东省电力公司滨州市滨城区供电公司 | Fault detection method, system, terminal and storage medium based on permutation entropy algorithm |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107241740A (en) * | 2017-04-25 | 2017-10-10 | 集美大学 | A kind of isolated island mobile power system health status monitoring sensor portion arranging method |
| CN110646685A (en) * | 2019-08-07 | 2020-01-03 | 国网浙江海盐县供电有限公司 | Comprehensive evaluation method for state of switch cabinet |
| CN112561226A (en) * | 2020-10-09 | 2021-03-26 | 国网四川省电力公司天府新区供电公司 | Novel high-voltage switch cabinet operation state comprehensive evaluation method |
| CN112800600A (en) * | 2021-01-18 | 2021-05-14 | 浙江工业大学 | Hidden attack detection method based on multi-scale sample entropy and Bayesian network |
-
2021
- 2021-06-11 CN CN202110654420.1A patent/CN113379252A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107241740A (en) * | 2017-04-25 | 2017-10-10 | 集美大学 | A kind of isolated island mobile power system health status monitoring sensor portion arranging method |
| CN110646685A (en) * | 2019-08-07 | 2020-01-03 | 国网浙江海盐县供电有限公司 | Comprehensive evaluation method for state of switch cabinet |
| CN112561226A (en) * | 2020-10-09 | 2021-03-26 | 国网四川省电力公司天府新区供电公司 | Novel high-voltage switch cabinet operation state comprehensive evaluation method |
| CN112800600A (en) * | 2021-01-18 | 2021-05-14 | 浙江工业大学 | Hidden attack detection method based on multi-scale sample entropy and Bayesian network |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113554131A (en) * | 2021-09-22 | 2021-10-26 | 四川大学华西医院 | Medical image processing and analyzing method, computer device, system and storage medium |
| CN114167838A (en) * | 2021-12-03 | 2022-03-11 | 西安电子科技大学 | Multi-scale health assessment and fault prediction method for servo system |
| CN114167838B (en) * | 2021-12-03 | 2023-07-07 | 西安电子科技大学 | Multi-scale health assessment and fault prediction method for servo system |
| CN117056086A (en) * | 2023-10-11 | 2023-11-14 | 国网山东省电力公司滨州市滨城区供电公司 | Fault detection method, system, terminal and storage medium based on permutation entropy algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101574615B1 (en) | A partial discharge monitoring and diagnosis system for power devices by using signal detection based on statistical phase-angle patterns | |
| CN113379252A (en) | Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy | |
| KR101553005B1 (en) | A partial discharge monitoring and diagnosis system for power devices | |
| Soualhi et al. | Hidden Markov models for the prediction of impending faults | |
| KR101574613B1 (en) | A detection and diagnosis system with remote configuration function for partial discharge by detecting UHF electrical signal | |
| EP3837565B1 (en) | Method, apparatus and device for evaluating the state of a distribution transformer, and a medium and a program | |
| Arabaci et al. | A knowledge-based diagnosis algorithm for broken rotor bar fault classification using FFT, principal component analysis and support vector machines | |
| CN116244617A (en) | Intelligent high-voltage switch cabinet fault diagnosis method and system based on heterogeneous graph structure learning | |
| Ferreira et al. | An application of machine learning approach to fault detection of a synchronous machine | |
| Mbamaluikem et al. | An artificial neural network-based intelligent fault classification system for the 33-kV Nigeria transmission line | |
| CN111999591B (en) | Method for identifying abnormal state of primary equipment of power distribution network | |
| US11693041B2 (en) | Method for monitoring the electric insulation status of a piece of equipment for MV or HV electric systems | |
| CN114397569A (en) | Circuit breaker fault arc detection method based on VMD parameter optimization and sample entropy | |
| CN116400244B (en) | Abnormality detection method and device for energy storage battery | |
| Mosallam et al. | Integrated bayesian framework for remaining useful life prediction | |
| CN115327325A (en) | Charging pile series arc fault detection method and device and electronic equipment | |
| CN115902557A (en) | Switch cabinet fault diagnosis processing method and device and nonvolatile storage medium | |
| Wilson et al. | Uncertainty Quantification of Capacitor Switching Transient Location Using Machine Learning | |
| CN112883639A (en) | GIS equipment service life prediction device and method based on machine learning | |
| CN112001326B (en) | Vibration fingerprint-based high-voltage switch equipment fault diagnosis and prediction method and system | |
| Tran et al. | Machine fault diagnosis and condition prognosis using classification and regression trees and neuro-fuzzy inference systems | |
| Zhong et al. | Failure prediction for linear ball bearings based on wavelet transformation and self-organizing map | |
| Van Tung Tran et al. | A classification and regression trees (CART) model of parallel structure and long-term prediction prognosis of machine condition | |
| Hu et al. | A fault prediction method based on IAALO-SVM and similarity measure | |
| Solis-Munoz et al. | A Novel Methodology for Power Quality Disturbances Detection and Classification in Industrial Facilities |
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 |