CN109523032A - Device parameter monitoring evidence network construction method based on independence analysis and test - Google Patents
Device parameter monitoring evidence network construction method based on independence analysis and test Download PDFInfo
- Publication number
- CN109523032A CN109523032A CN201811470575.4A CN201811470575A CN109523032A CN 109523032 A CN109523032 A CN 109523032A CN 201811470575 A CN201811470575 A CN 201811470575A CN 109523032 A CN109523032 A CN 109523032A
- Authority
- CN
- China
- Prior art keywords
- node
- value
- model
- equipment
- input data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 57
- 238000012360 testing method Methods 0.000 title claims abstract description 20
- 238000010276 construction Methods 0.000 title claims description 10
- 238000000034 method Methods 0.000 claims abstract description 25
- 238000012417 linear regression Methods 0.000 claims description 5
- 238000002955 isolation Methods 0.000 claims 1
- 238000005070 sampling Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 9
- 230000004927 fusion Effects 0.000 description 9
- 230000001364 causal effect Effects 0.000 description 5
- 239000000110 cooling liquid Substances 0.000 description 3
- 239000000446 fuel Substances 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 241001061076 Melanonus zugmayeri Species 0.000 description 1
- 235000013290 Sagittaria latifolia Nutrition 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 235000015246 common arrowhead Nutrition 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
- F02C—GAS-TURBINE PLANTS; AIR INTAKES FOR JET-PROPULSION PLANTS; CONTROLLING FUEL SUPPLY IN AIR-BREATHING JET-PROPULSION PLANTS
- F02C9/00—Controlling gas-turbine plants; Controlling fuel supply in air- breathing jet-propulsion plants
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F02—COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
- F02B—INTERNAL-COMBUSTION PISTON ENGINES; COMBUSTION ENGINES IN GENERAL
- F02B77/00—Component parts, details or accessories, not otherwise provided for
- F02B77/08—Safety, indicating, or supervising devices
Landscapes
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The method for constructing the equipment parameter monitoring evidence network based on the independence analysis and test is characterized in that different equipment parameters of target equipment are synchronously acquired by utilizing a plurality of sensors, and each equipment parameter is used as a node in the parameter monitoring evidence network of the target equipment. And synchronously sampling by each sensor at the same time interval in the running time of the target equipment, and taking the monitoring data sampled by each sensor in the running time of the target equipment as the node input data of each node. And constructing a undirected evidence network structure model based on the maximum information coefficient among the nodes. And finally, determining the direction of the undirected edge according to the noise model on the basis of the undirected evidence network model, and constructing a device parameter monitoring evidence network corresponding to the target device. According to the invention, historical data of a plurality of different device parameters of the target device are detected by using the sensor, and the network structure is constructed based on the historical data, so that the objectivity and the accuracy of the evidence network structure are improved.
Description
Technical Field
The invention relates to the technical field of multivariate information fusion in equipment monitoring, in particular to an equipment parameter monitoring evidence network construction method based on independence analysis and test.
Background
In the multivariate information fusion in equipment monitoring, data relevance and information uncertainty are ubiquitous. In order to comb the relationship between the multiple information and further realize the effective fusion of the multiple information, the relationship between a plurality of mutually related and mutually influenced different types of information needs to be considered, and the uncertainty caused by the combined action of various factors existing inside and outside in the process of acquiring the parameter information of the equipment is needed. If the uncertainties cannot be correctly described, the mutual relationship between the information cannot be reasonably described, and further the effective fusion of the multivariate information cannot be realized.
The evidence network is used as organic combination of a graph model and an evidence theory, and has the capability of expressing complex association relations by the graph model and the advantage of describing uncertain knowledge by the evidence theory. The expert system expressed according to the network structure can qualitatively and quantitatively describe the causal relationship among different information and make reasoning according to corresponding observation data. Therefore, the evidence network is one of the most effective theoretical models in the fields of complex uncertain knowledge expression and reasoning at present, and is widely applied to complex system management decision practices such as weapon equipment system capability evaluation, reliability evaluation, state monitoring, medical diagnosis and the like.
The information expression of the evidence network consists of two parts: one part is a network structure represented by a directed acyclic graph, each node in the network represents one parameter in actual sample data, the connection among the nodes represents the causal relationship among variables, the node pointed by an arrow tail represents a reason, and the node pointed by the arrow head represents a result; the other part is evidence network parameters, or confidence rule base, which expresses the influence degree of the reason to the result. When the problem of the multi-element information fusion in equipment monitoring is solved, the established evidence network structure is the parameter network related to the multi-element information fusion problem.
The graph model is used as a structural model of the evidence network, describes the incidence relation among variables, and is the basis for analyzing and solving problems by applying an evidence network reasoning method. If the incidence relation between the variables is omitted or misjudged by the evidence network structure model, the network reasoning result is influenced, and the reliability of analyzing and solving the problem by the evidence network model is further reduced.
At present, methods for directly constructing network structure models according to expert experience knowledge are mostly adopted in evidence network related application research. Although the method is simple and rapid, the evidence network structure built only by relying on the domain knowledge has certain defects and limitations, the built network model is not high in accuracy, the relation and mutual influence among equipment parameters in equipment monitoring cannot be objectively reflected, and the accuracy of the result analyzed and inferred based on the network model is not high naturally.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a device parameter monitoring evidence network construction method based on independence analysis and test. With the improvement of sensor sampling technology and database storage capacity, a large amount of relevant historical data is available in the solution analysis of practical problems. According to the method, the historical data of a plurality of different equipment parameters of the target equipment are acquired by using the sensor, the equipment parameter monitoring evidence network is constructed based on the historical data, the objectivity of the constructed evidence network structure is improved, and the accuracy of the reasoning analysis result of the evidence network in the multivariate information fusion problem is improved.
In order to realize the purpose of the invention, the following technical scheme is adopted for realizing the purpose:
referring to fig. 4, the method for constructing the equipment parameter monitoring evidence network based on the independence analysis and test includes the following steps:
s1, determining target equipment, and collecting k equipment parameters of the target equipment by using k (k is larger than 1) sensors respectively, wherein each equipment parameter is used as a node in a parameter monitoring evidence network of the target equipment, namely, k nodes are provided.
The method comprises the steps that each sensor synchronously samples within the running time of target equipment according to the same time interval, monitoring data sampled within the running time of the target equipment by each sensor is used as node input data of each node, and if each sensor samples for n times within the running time of the target equipment, the node input data of each node comprises n sample data. The collection of node input data for all nodes is denoted as data set D.
S2, constructing the undirected evidence network structure model based on the maximum information coefficient.
S2.1, calculating the maximum information coefficient between each node in the data set D and other nodes in the data set D.
For any node x in the data set DiCalculating node xiWith any other node x in the data set DjThe maximum information coefficient in between, the method is as follows:
node xiThe node of (1) sets all values within the range of the minimum value and the maximum value in the input data as a node xiThe value interval of (2), the node xjThe node of (1) sets all values within the range of the minimum value and the maximum value in the input data as a node xjThe value range of (2). Node xiIs divided into a sub-intervals, and the node x is divided into a sub-intervalsjThe value interval is divided into b sub-intervals, the value interval can be arbitrarily divided under the condition that the constraint ab < B (n) is met, so that the aim of maximizing the MIC value is fulfilled, the division of the value interval is not required to be uniform, and the division mode is marked as G. Node xiIs tabulated as Representing a node xiAnd node xjThe maximum information coefficient in between, has symmetry.
Where the variable n represents the number of sample sets and the function b (n) limits the size of the search grid. According to experimental verification, B (n) is set as B (n) ═ n0.6It is more suitable. In the present invention, B (n) is n0.6。I(D|G) And a mutual information value representing the data set D divided by G. Let node xiThe joint distribution of the nodes j and under the division of G is p (x)i,xj) Marginal distribution is p (x) respectivelyi) And p (x)j) Then the mutual information value can be calculated as follows;
in the same way as above, node x is computediObtaining the maximum information coefficient between the node x and other nodes in the data set DiList of maximum information coefficients
S2.2, correcting the maximum information coefficient list among the nodes in the data set D.
For any node x in the data set DiThe corrected maximum information coefficient is listed as
To avoid introducing redundant edgesThe generation of the triangular ring needs to correct the maximum information coefficient of each node. In the invention, a penalty factor delta is introduced, and the larger the delta is, the more difficult redundant edges are introduced, but the possibility of missing edges in an actual evidence network structure is increased; the smaller the δ, the less likely it is to miss edges present in the actual evidence network, but the greater the likelihood of introducing redundant edges. The value of the penalty factor delta is determined according to empirical knowledge, and generally delta is selected to be E [0, 0.5 ∈]. Node h represents a common occurrence at node xiList of maximum information coefficientsMiddle node xjFront and node xjList of maximum information coefficientsMiddle node xiThe previous class of nodes. m represents the total number of such nodes.
S2.3, judging whether an undirected edge exists between the nodes.
S2.3.1 modified maximum information coefficient list for all nodesIn descending order, the maximum value is recorded as
S2.3.2 setting an isolated threshold factor gamma, the value of which is determined according to empirical knowledge, and the value of gamma is 0, 0.3]. If it is notThen node xiFor an orphaned node, there is no undirected edge, otherwise, S2.3.3 is performed;
s2.3.3 if node xiAnd node xjThe corrected maximum information coefficient satisfies:
then node xiAnd node xjThere is no side between them, wherein α is connection threshold factor, α is bigger, connection condition is more rigorous, it is not easy to introduce redundant side, α is smaller, connection condition is more loose, it is not easy to cause omission, the value of connection threshold factor α is determined according to empirical knowledge, generally α is from [0.7, 0.9 ]]。
S3, based on the undirected evidence network model, determining the undirected direction according to the noise model, and constructing a device parameter monitoring evidence network corresponding to the target device.
S3.1 at node xiAnd node xjRespectively constructing nonlinear regression models x on the premise of existence of undirected edgesi:＝f(xj)+n1And xj:＝f(xi)+n2Model xi:＝f(xj)+n1F (x) of (1)j) Description node x obtained by finger fittingjFor reasons, node xiAs a resulting nonlinear regression function, n1Finger model xi:＝f(xj)+n1The noise term of (1); model xj:＝f(xi)+n2F (x) of (1)i) Description node x obtained by finger fittingiFor reasons, node xjAs a resulting nonlinear regression function, n2Finger model xj:＝f(xi)+n2The noise term in (1).
S3.2 calculating residual n of two models in S3.1 respectively1＝xi-f(xj) And n2＝xj-f(xi)，f(xj) Finger node xjBy substituting the node input data into a non-linear regression model xi:＝f(xj)+n1Node x obtained lateriValue of (a), f (x)i) Finger node xiBy substituting the node input data into a non-linear regression model xj:＝f(xi)+n2Node x obtained laterjThe numerical value of (c).
S3.3 Pair xiAnd n1、xjAnd n2The two groups of variables were subjected to T test, respectively. By calculating test statisticsAndobtaining the corresponding t value, which is recorded as t1And t2Looking up the T test statistical table to obtain p values, and respectively recording the obtained p values as p1And p2. Wherein,representing a node xjMean, σ, of the corresponding node input data1Representing a node xjThe standard deviation of the input data of the corresponding node,representing a node xiMean, σ, of the corresponding node input data2Representing a node xiThe standard deviation of the corresponding node input data, and n represents the number of samples of the node input data collected by each node in the data set D.
S3.4 comparison of p1And p2If p is1＞p2Then accept model xi:＝f(xj)+n1Consider node xiAnd node xjThe direction without the edge therebetween is the slave node xjPointing to node xi(ii) a If p is2＞p1Then accept model xj:＝f(xi)+n2Consider node xjAnd node xiThe direction without the edge therebetween is the slave node xiPointing to node xj；
S3.5 determines the direction of all the non-directional edges according to S3.1 to S3.4, on the basis of which it is checked whether a loop exists in the directed graph. If a loop exists, the edge with the smallest p value is deleted. And thus, constructing the equipment parameter monitoring evidence network corresponding to the target equipment.
The method takes the maximum information coefficient between each pair of nodes in the data set as the description of the strength of the causal relationship between the nodes, and the larger the maximum information coefficient is, the stronger the causal relationship between the nodes is, the more likely the edges exist between the nodes, and vice versa. In order to avoid the generation of redundant edges, a penalty factor is introduced to correct the maximum information coefficient. And designing a connection rule to analyze the corrected maximum information coefficient so as to determine a undirected evidence network structure. Further, on the basis of generating the undirected evidence network structure, the directed network structure, i.e. the direction of the edges, is determined. And (4) respectively carrying out causal relationship analysis on each undirected edge by using a noise adding model, and determining the direction of the edge.
The invention has the following beneficial effects:
(1) for a target object, namely target equipment, a plurality of sensors are used for synchronously acquiring different equipment parameters of the target equipment, and historical data of the different equipment parameters of the target equipment are used for constructing a network structure, so that the objectivity of the evidence network structure is improved, and the accuracy of the reasoning analysis result of the evidence network in the multivariate information fusion problem is improved. The relationship and the mutual influence between different device parameters of the target device can be effectively described from the constructed network structure.
(2) The invention adopts the maximum information coefficient to measure the correlation degree between the nodes. As a coefficient for identifying the correlation between variables, the maximum information coefficient can be accurately identified regardless of whether the relationship between the variables is a functional relationship or a non-functional relationship. By such properties, a more reliable undirected network structure can be obtained in the end.
(3) The invention adopts a noise adding model to determine the direction of the edge in the network, thereby obtaining a directed network structure. The noise adding model is a mature direction determining method, and can effectively determine the direction of edges between nodes by using data. Through the mechanism, the direction of each edge is gradually determined, and finally a directional evidence network structure is formed.
Drawings
FIG. 1 is a diagram of an evidence network for monitoring three engine parameters to be established in an embodiment; wherein FIG. 1(1) is a diagram of an evidence network architecture for monitoring turbine engine parameters to be established; FIG. 1(2) is a diagram of an evidence network illustrating the monitoring of horizontal engine parameters to be established; FIG. 1(3) is a diagram of an evidence network for monitoring of rotor engine parameters to be established.
FIG. 2 is a model of a undirected network of turbine engine parameter monitoring constructed using the method of the present invention;
FIG. 3 is a diagram of an evidence network for turbine engine parameter monitoring constructed using the method of the present invention.
FIG. 4 is a flow chart of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the drawings of the embodiments of the present invention, and further detailed description will be given, but the embodiments of the present invention are not limited thereto.
The following three engines are respectively used as target equipment, and equipment parameter monitoring evidence networks of the three engines are respectively constructed by utilizing the equipment parameter monitoring evidence network construction method based on independence analysis and test provided by the invention. The three engines are respectively a turbine engine, a horizontal engine and a rotor engine.
FIG. 1 is a diagram of an evidence network for three engine parameter monitoring to be established, wherein:
fig. 1(1) shows a diagram of an evidence network for monitoring parameters of a turbine engine to be established, node a showing a flow rate of fuel at an outlet of a high-pressure compressor of the turbine engine and a static pressure ratio, node B showing a flow rate of cooling liquid of a high-pressure turbine of the turbine engine, and node C showing an enthalpy of exhaust gas of the turbine engine.
Fig. 1(2) shows an evidence network structure diagram of parameter monitoring of a horizontal engine to be established, wherein a node A shows the outlet temperature of a low-pressure compressor of the horizontal engine, a node B shows the outlet temperature of the high-pressure compressor of the horizontal engine, and a node C shows the outlet temperature of a low-pressure turbine of the horizontal engine.
Fig. 1(3) shows a structure diagram of an evidence network for monitoring parameters of a rotary engine to be established, wherein node a shows a combustion air ratio of a combustor of the rotary engine, node B shows a core rotation speed of the rotary engine, and node C shows a bypass pressure of the rotary engine.
In order to verify the feasibility and the effectiveness of the method for constructing the equipment parameter monitoring evidence network based on the independence analysis and test, 200 groups of monitoring data are collected for each of the three engine parameters related in the figure 1, and the method provided by the invention is adopted to construct the equipment parameter monitoring evidence network of the three engines. The steps of the present invention will be described by taking the turbine engine in fig. 1(1) as an example:
s1, determining that the target equipment is a turbine engine, and synchronously sampling the flow rate and static pressure ratio of the fuel at the outlet of the high-pressure compressor, the flow rate of the cooling liquid of the high-pressure turbine and the exhaust enthalpy according to the same time interval in the operating life cycle of the turbine engine by using the sensors.
The 'ratio of the fuel flow and the static pressure at the outlet of the high-pressure compressor' is recorded as a node A, the flow rate of the high-pressure turbine cooling liquid 'is recorded as a node B, and the exhaust enthalpy' is recorded as a node C. And the monitoring data sampled by each sensor is used as node input data of each node. The collection of node input data for all nodes is denoted as data set D.
S2, constructing a undirected evidence network structure model based on the maximum information coefficient:
s2.1, calculating the maximum information coefficient between the nodes in pairs:
TABLE 1 maximum information coefficient between nodes
| Maximum information coefficient | A | B | C |
| A | 1 | 1 | 0.8596 |
| B | 1 | 1 | 0.8778 |
| C | 0.8596 | 0.8778 | 1 |
The maximum information coefficient list of each node is as in tables 2 to 4:
table 2 maximum information coefficient list for node a
Table 3 maximum information coefficient list of node B
| Maximum information coefficient list | MICB |
| A | 1 |
| B | 1 |
| C | 0.8778 |
TABLE 4 maximum information coefficient List for node C
| Maximum information coefficient list | MICC |
| C | 1 |
| B | 0.8778 |
| A | 0.8596 |
S2.2, correcting the maximum information coefficient among the nodes;
taking the maximum information coefficient between the node a and the node C as an example, if δ is 0.1, then:
TABLE 5 maximum information coefficient between nodes after correction
S2.3, judging whether an undirected edge exists between the nodes:
S2.3.1
s2.3.2 setting the isolated threshold factor γ to 0.3, it is clear from the results S2.3.1 that there are no isolated nodes;
s2.3.3 sets the connection threshold factor α to 0.8 and checks whether the modified maximum information coefficient between nodes satisfies:
if so, a non-directional edge is generated. It is verified that there is a non-directional edge between node A and node B, and between node B and node C. As shown in fig. 2.
S3, based on the undirected evidence network model, determining the direction of the undirected edge according to the noise model: (analysis is performed by taking node A and node B as examples)
S3.1, non-directional edges exist between the node A and the node B, and non-linear regression models are respectively constructed:
A:＝-0.0062B2+0.927B +19.553 and B ═ 0.0127A2+3.3923A-76.891；
S3.2 calculating residual n of two models in step S3.1 respectively1A-f (b) and n2＝B-f(A)；
S3.3 Pair xiAnd n1、xjAnd n2And (3) respectively carrying out T test on the two groups of variables to obtain p values as follows: p is a radical of1＝2.5×10-198And p2＝5.06×10-67；
S3.4 comparison of p1And p2Due to p2＞p1Therefore, model B: ═ f (A) + n is accepted2Considering that the direction without a side between the node A and the node B is from the node A to the node B;
and S3.5, determining that the direction of the edge between the node B and the node C is from the node B to the node C according to S3.1-S3.4, and judging that no loop exists in the network. The network structure obtained at this time is an evidence network for monitoring the turbine engine parameters in fig. 1(1), as shown in fig. 3.
The evidence network for monitoring the parameters of the horizontal engine in the graph 1(2) and the evidence network for monitoring the parameters of the rotary engine in the graph 1(3) are respectively constructed according to the same method steps for constructing the evidence network for monitoring the parameters of the turbine engine, and the results are shown in a table 6:
table 6 evidence network structure learning results
Table 6 shows the results of the network structure learning of the evidence for monitoring three engine parameters according to the present invention. The fourth column "redundant edge" in table 6 describes the number of edges that do not exist in the evidence network structure for device parameter monitoring to be established but exist in the evidence network structure for device parameter monitoring that is finally constructed by the method of the present invention; the fifth column "missing edges" describes the number of edges that exist in the evidence network structure of device parameter monitoring to be established and do not exist in the evidence network structure of device parameter monitoring that is finally established by the method of the present invention. As can be seen from Table 6, the evidence network structure for monitoring the equipment parameters can be effectively and accurately constructed and obtained on the basis of small sample data. In the evidence network structure for monitoring the three engine equipment parameters, which is constructed in the example, redundant edges and missing edges do not exist. For three engine parameter monitoring network structures in the example, the parameter monitoring network structure can be accurately constructed by using 200 groups of sample data for each type of engine. The invention solves the problem that the network structure is constructed by depending on expert experience knowledge in the current evidence network modeling process, provides the data-driven evidence network structure learning method, greatly improves the objectivity of the evidence network structure, and further improves the reliability of the evidence network reasoning result, so the method has high utilization value and good application prospect in solving the problem of multivariate information fusion.
In summary, although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made by those skilled in the art without departing from the spirit and scope of the invention.
Claims (5)
1. The equipment parameter monitoring evidence network construction method based on independence analysis and test is characterized by comprising the following steps of:
s1, determining target equipment, and collecting k equipment parameters of the target equipment by using k (k is more than 1) sensors respectively, wherein each equipment parameter is used as a node in a parameter monitoring evidence network of the target equipment, namely k nodes are provided;
the method comprises the steps that each sensor synchronously samples within the running time of target equipment according to the same time interval, monitoring data sampled within the running time of the target equipment by each sensor is used as node input data of each node, and if each sensor samples for n times within the running time of the target equipment, the node input data of each node comprises n sample data; the set of node input data of all nodes is recorded as a data set D;
s2, constructing a undirected evidence network structure model based on the maximum information coefficient;
s2.1, calculating the maximum information coefficient between each node in the data set D and each other node in the data set D;
s2.2, correcting the maximum information coefficient list among the nodes in the data set D;
s2.3, judging whether undirected edges exist between the nodes;
s3, based on the undirected evidence network model, determining the undirected direction according to the noise model, and constructing a device parameter monitoring evidence network corresponding to the target device.
2. The independence analysis and testing based equipment parameter monitoring evidence network construction method according to claim 1, characterized in that in S2.1 for any node x in the data set DiCalculating node xiWith any other node x in the data set DjThe maximum information coefficient in between, the method is as follows:
node xiThe node of (1) sets all values within the range of the minimum value and the maximum value in the input data as a node xiThe value interval of (2), the node xjThe node of (1) sets all values within the range of the minimum value and the maximum value in the input data as a node xjThe value range of (1); node xiIs divided into a sub-intervals, and the node x is divided into a sub-intervalsjThe value interval is divided into b sub-intervals, the division of the value interval is not required to be uniform, the value interval can be arbitrarily divided under the condition that the constraint ab < B (n) is met, and the division mode is marked as G; node xiIs tabulated as Representing a node xiAnd node xjThe maximum information coefficient in between, has symmetry;
where the variable n represents the number of sample sets, the function b (n) limits the size of the search grid, and b (n) n0.6；I(D|G) Representing the mutual information value of the data set D with the division mode G, and setting a node xiAnd nodejWith a joint distribution under the division of G of p (x)i,xj) Marginal distribution is p (x) respectivelyi) And p (x)j) Then the mutual information value can be calculated as follows;
in the same way as above, node x is computediObtaining the maximum information coefficient between the node x and other nodes in the data set DiList of maximum information coefficients
3. The independence analysis and testing based equipment parameter monitoring evidence network construction method according to claim 2, characterized in that in S2.2, for any node x in the data set DiThe corrected maximum information coefficient is listed as
Wherein, delta is a penalty factor, and is delta epsilon [ c ], [ alpha ]0，0.5](ii) a Node h represents a common occurrence at node xiList of maximum information coefficientsMiddle node xjFront and node xjList of maximum information coefficientsMiddle node xiThe previous class of nodes, m, represents the total number of such nodes.
4. The independence analysis and test based equipment parameter monitoring evidence network construction method according to claim 3, wherein the S2.3 implementation method is as follows:
s2.3.1 modified maximum information coefficient list for all nodesIn descending order, the maximum value is recorded as
S2.3.2 setting an isolation threshold factor gamma, taking gamma as 0, 0.3](ii) a If it is notThen node xiFor an orphaned node, there is no undirected edge, otherwise, S2.3.3 is performed;
s2.3.3 if node xiAnd node xjThe corrected maximum information coefficient satisfies:
then node xiAnd node xjThere is no directional edge between them, where α is the connection threshold factor, and α E is taken as [0.7, 0.9 ]]。
5. The independence analysis and test based equipment parameter monitoring evidence network construction method according to claim 4, wherein the implementation method of S3 is as follows:
s3.1 at node xiAnd node xjRespectively constructing nonlinear regression models x on the premise of existence of undirected edgesi:＝f(xj)+n1And xj:＝f(xi)+n2Model xi:＝f(xj)+n1F (x) of (1)j) Description node x obtained by finger fittingjFor reasons, node xiAs a resulting nonlinear regression function, n1Finger model xi:＝f(xj)+n1The noise term of (1); model xj:＝f(xi)+n2F (x) of (1)i) Description node x obtained by finger fittingiFor reasons, node xjAs a resulting nonlinear regression function, n2Finger model xj:＝f(xi)+n2The noise term of (1);
s3.2 calculating residual n of two models in S3.1 respectively1＝xi-f(xj) And n2＝xj-f(xi)，f(xj) Finger node xjBy substituting the node input data into a non-linear regression model xi:＝f(xj)+n1Node x obtained lateriValue of (a), f (x)i) Finger node xiBy substituting the node input data into a non-linear regression model xj:＝f(xi)+n2Node x obtained laterjThe value of (d);
s3.3 Pair xiAnd n1、xjAnd n2Carrying out T test on the two groups of variables respectively; by calculating test statisticsAndobtaining the corresponding t value, which is recorded as t1And t2Value, recheckReading the T test statistical table to obtain p values, and respectively recording the obtained p values as p1And p2(ii) a Wherein,representing a node xjMean, σ, of the corresponding node input data1Representing a node xjThe standard deviation of the input data of the corresponding node,representing a node xiMean, σ, of the corresponding node input data2Representing a node xiThe standard deviation of the corresponding node input data, n represents the sample number of the node input data collected by each node in the data set D;
s3.4 comparison of p1And p2If p is1＞p2Then accept model xi:＝f(xj)+n1Consider node xiAnd node xjThe direction without the edge therebetween is the slave node xjPointing to node xi(ii) a If p is2＞p1Then accept model xj:＝f(xi)+n2Consider node xjAnd node xiThe direction without the edge therebetween is the slave node xiPointing to node xj；
S3.5, determining the directions of all the non-directional edges according to S3.1 to S3.4, and checking whether a loop exists in the directed graph or not on the basis; if the loop exists, deleting the edge with the minimum p value; and thus, constructing the equipment parameter monitoring evidence network corresponding to the target equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811470575.4A CN109523032B (en) | 2018-12-04 | 2018-12-04 | Device parameter monitoring evidence network construction method based on independence analysis and test |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811470575.4A CN109523032B (en) | 2018-12-04 | 2018-12-04 | Device parameter monitoring evidence network construction method based on independence analysis and test |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109523032A true CN109523032A (en) | 2019-03-26 |
| CN109523032B CN109523032B (en) | 2021-09-17 |
Family
ID=65793900
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811470575.4A Active CN109523032B (en) | 2018-12-04 | 2018-12-04 | Device parameter monitoring evidence network construction method based on independence analysis and test |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109523032B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950850A (en) * | 2020-07-10 | 2020-11-17 | 北京航空航天大学 | Evidence network-based unmanned aerial vehicle system guarantee capability evaluation method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3036875A1 (en) * | 2013-08-20 | 2016-06-29 | Nokia Technologies Oy | A method and apparatus for privacy-enhanced evidence evaluation |
| CN107563596A (en) * | 2017-08-03 | 2018-01-09 | 清华大学 | A kind of evaluation index equilibrium state analysis method based on Bayes's causal network |
| CN108830382A (en) * | 2018-05-31 | 2018-11-16 | 上海电力学院 | A kind of DCS method for evaluating reliability based on dynamic evidence network |
-
2018
- 2018-12-04 CN CN201811470575.4A patent/CN109523032B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3036875A1 (en) * | 2013-08-20 | 2016-06-29 | Nokia Technologies Oy | A method and apparatus for privacy-enhanced evidence evaluation |
| CN107563596A (en) * | 2017-08-03 | 2018-01-09 | 清华大学 | A kind of evaluation index equilibrium state analysis method based on Bayes's causal network |
| CN108830382A (en) * | 2018-05-31 | 2018-11-16 | 上海电力学院 | A kind of DCS method for evaluating reliability based on dynamic evidence network |
Non-Patent Citations (3)
| Title |
|---|
| 姜江: "证据网络建模、推理及学习方法研究", 《中国优秀硕士学位论文全文数据库基础科学辑》 * |
| 李中杰: "基于证据网络的多源敌我识别信息融合处理方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 郭强: "一种基于区间规则的条件证据网络推理决策方法", 《控制与决策》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950850A (en) * | 2020-07-10 | 2020-11-17 | 北京航空航天大学 | Evidence network-based unmanned aerial vehicle system guarantee capability evaluation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109523032B (en) | 2021-09-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109522600B (en) | Complex equipment residual service life prediction method based on combined deep neural network | |
| CN108664402B (en) | Fault prediction method based on software network feature learning | |
| Lindemann et al. | Anomaly detection and prediction in discrete manufacturing based on cooperative LSTM networks | |
| CN108256173B (en) | Gas circuit fault diagnosis method and system for dynamic process of aircraft engine | |
| Fu et al. | A re-optimized deep auto-encoder for gas turbine unsupervised anomaly detection | |
| CN109635318B (en) | Intelligent analysis redundancy design method for aero-engine sensor | |
| CN109522221B (en) | Method and system for improving fuzzy test efficiency | |
| CN111637045B (en) | Fault diagnosis method for air compressor of ocean platform | |
| CN114492533B (en) | Construction method and application of variable-working-condition bearing fault diagnosis model | |
| Xiong et al. | Adaptive deep learning-based remaining useful life prediction framework for systems with multiple failure patterns | |
| CN115081484A (en) | Aircraft engine sensor fault diagnosis method based on CRJ-OSELM algorithm | |
| Jiang et al. | Remaining useful life estimation combining two-step maximal information coefficient and temporal convolutional network with attention mechanism | |
| CN109523032B (en) | Device parameter monitoring evidence network construction method based on independence analysis and test | |
| Duan et al. | Fault diagnosis of air compressor in nuclear power plant based on vibration observation window | |
| CN105228185A (en) | A kind of method for Fuzzy Redundancy node identities in identification communication network | |
| Zhang et al. | Hierarchical hybrid testability modeling and evaluation method based on information fusion | |
| CN117454145A (en) | Aeroengine fault diagnosis method based on multi-mode sensor | |
| CN116737529A (en) | Multidimensional software data reliability verification method | |
| CN111753876B (en) | Product quality detection method based on deep neural network | |
| CN110750756B (en) | Real-time on-line instrument checksum diagnosis method through optimal support vector machine algorithm | |
| Liu et al. | Semi-Supervised Dynamic Latent Variable Regression for Prediction and Quality-Relevant Fault Monitoring | |
| CN115879355A (en) | Temperature compensation method of piezoelectric sensor | |
| Chen et al. | A Multivariate Time Series Anomaly Detection Method Based on Generative Model | |
| Gao et al. | Research on fault diagnosis of electric gate valve in nuclear power plant based on the VMD-MDI-ISSA-RF model | |
| Zhou et al. | A novel information fusion model based on DS evidence theory for equipment diagnosis |
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 |