CN104503441A - Process fault monitoring method based on improved dynamic visible graph - Google Patents

Process fault monitoring method based on improved dynamic visible graph Download PDF

Info

Publication number
CN104503441A
CN104503441A CN201410806907.7A CN201410806907A CN104503441A CN 104503441 A CN104503441 A CN 104503441A CN 201410806907 A CN201410806907 A CN 201410806907A CN 104503441 A CN104503441 A CN 104503441A
Authority
CN
China
Prior art keywords
mrow
msub
alpha
data
network
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
Application number
CN201410806907.7A
Other languages
Chinese (zh)
Other versions
CN104503441B (en
Inventor
耿志强
王尊
朱群雄
韩永明
徐圆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Chemical Technology
Original Assignee
Beijing University of Chemical Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing University of Chemical Technology filed Critical Beijing University of Chemical Technology
Priority to CN201410806907.7A priority Critical patent/CN104503441B/en
Publication of CN104503441A publication Critical patent/CN104503441A/en
Application granted granted Critical
Publication of CN104503441B publication Critical patent/CN104503441B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00Testing or monitoring of control systems or parts thereof
    • G05B23/02Electric testing or monitoring
    • G05B23/0205Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0218Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
    • G05B23/0224Process history based detection method, e.g. whereby history implies the availability of large amounts of data
    • G05B23/0227Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention provides a process fault monitoring method based on an improved dynamic visible graph. With the help of a complicated network theory, the invention provides an improved dynamic visible graph algorithm, on the basis of the algorithm, time serial data is mapped into a complicated network structure, the time serial data with different variables can be distinguished and identified through network characteristics, in addition, whether production data generates a fault or not is judged. The method has the advantages that the fault false alarm rate and the missing report rate be reduced, in addition, the fault occurrence can be earlier monitored, and the real-time monitoring of the complicated industrial process can be more favorably realized.

Description

Process fault monitoring method based on improved dynamic visible graph
Technical Field
The invention relates to the field of fault identification, in particular to a process fault monitoring method based on an improved dynamic visible graph (MDVG).
Background
At present, the rapid development of information technology provides great convenience for acquiring and processing mass data, and a complex network is used as a new research field to attract the wide attention of numerous disciplines. In recent years, a lot of efforts have been made to extract relevant information from data generated by complex systems in various fields, describe and analyze the attribute states of the systems using network knowledge. The time sequence data are mapped into the complex network, and the complex time sequence data are analyzed by applying rich and advanced complex network analysis methods, which is particularly attractive.
Unlike the relatively conventional distance-based and correlation coefficient-based network construction methods, Lacasa et al creatively propose a visibility algorithm, namely, Natural Visibility Graph (NVG). And corresponding each data in the time sequence data to a node of the complex network, and connecting two nodes if other nodes meet the visible condition of each node, and forming a visible graph. The constructed network inherits the attribute characteristics of the original sequence on the structure, such as periodic, random and fractal sequences can be respectively converted into regular, random and scale-free networks.
Subsequently, Fioriti et al proposed a Horizontal Visibility Graph (HVG) algorithm that calculates the maximum eigenvalue of the HVG-related adjacency matrix derived from multiple time series to distinguish between sequence chaos and randomness. The HVG is a subgraph of the NVG, which translates a set of time-series data into a unique network structure. Time series of complex structures associated with various phenomena, such as flow fluctuations, stock indices, heartbeat dynamics, random and chaotic sequences, etc., can be described and explored using NVG and HVG algorithms.
Based on the NVG algorithm, Bezsudnov et al propose a Dynamic Visibility Graph (DVG) algorithm, which changes the visible conditions by introducing a "view" parameter, thereby affecting the change of the network structure, and converts a time series into a group of networks, each DVG being a sub-graph of the NVG. Meanwhile, through three network characteristics of relative average degree of nodes, relative average connection length and number of unconnected groups, a new dynamic dimension is provided for distinguishing, identifying and describing various time sequences in detail
On the other hand, process monitoring methods can be mainly classified into three categories: model-based methods, knowledge-based methods, and data-based methods. Data-based process monitoring methods are widely used because they do not require process models, particularly in complex industrial processes or systems, such as chemical processes, where models and expert knowledge are difficult to build and acquire in practice, as compared to the first two more traditional methods. Due to the widespread use of distributed control systems in modern industrial processes, large amounts of data can be recorded and collected. The process data generally has the characteristics of high dimension, nonlinearity, time variation, multi-mode, autocorrelation and the like, and the existing process monitoring method based on the data is difficult to completely process.
Therefore, there is a need to develop a new data-based process detection method, so as to solve the above-mentioned drawbacks of the prior art.
Disclosure of Invention
In order to solve the above problems, the present invention proposes a process fault monitoring method based on an improved dynamic visibility graph (MDVG).
The invention provides a process fault monitoring method based on an improved dynamic visible diagram, which provides an improved dynamic visible diagram algorithm by means of a complex network theory, maps time sequence data to a complex network structure on the basis of the algorithm, distinguishes and identifies the time sequence data of different variables through network characteristics, and judges whether production data has faults or not.
The invention requests to protect a process fault monitoring method based on an improved dynamic visible diagram, which comprises the following steps:
s101, determining monitoring variables, respectively normalizing historical data of the variables according to a certain moving pane length, and mapping the variables to a complex network by using an MDVG algorithm;
the process of mapping to a complex network using the MDVG algorithm is as follows:
consider any two data (t) in a set of time series datai,xi) And (t)k,xk),i<k, for all data (t) between themj,xj),i<j<k, if the visible condition is satisfied
x j < x i + ( x k - x i ) t j - t i t k - t i - - - ( 1 )
<math> <mrow> <mi>&alpha;</mi> <mrow> <mo>(</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>&lt;</mo> <msub> <mi>&alpha;</mi> <mi>ik</mi> </msub> <mo>=</mo> <mi>arctan</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> </mrow> <mrow> <msub> <mi>t</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> </mrow> </mfrac> <mo>*</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow> </math>
Then, the nodes to which the two data maps are considered to be visible and connected in the network,
and (3) determining the corresponding unique network structure of the time series data under the viewing angle alpha and the time interval constant h through the calculation of the formulas (1) and (3).
S102, calculating three characteristic parameters K (alpha), Lambda (alpha) and Q (alpha) of the mapped network, and determining a monitoring index and a corresponding threshold value.
Three important characteristic parameters characterizing MDVG (α) are as follows:
(1) node relative average degree K (α):
<math> <mrow> <mi>K</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average of the network nodes at the viewing angles alpha and pi, respectively.
(2) Relative average connection length Λ (α):
<math> <mrow> <mi>&Lambda;</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average link lengths of the network at viewing angles alpha and pi, respectively.
(3) Number of non-connected groups Q (α): if a group of nodes arranged in sequence has at least one pair of connection relations between each node and any node in the non-group has no connection relation, the group of nodes forms a non-connected group. The number of the non-connected groups in each network is the number of the non-connected groups.
S103, carrying out online process monitoring, monitoring the current data of each variable by adopting the length of a movable pane which is the same as the historical data, and calculating the monitoring index of the current data;
s104, judging whether the monitoring index of the current data exceeds a threshold value, and if so, judging that the system is in a state of exceeding
And sending an alarm so that an operator can find and determine the fault reason in time.
Further, the time interval constant h may be determined by a particle swarm optimization algorithm, so that the mean of the occurrence times of the modes of K (α), Λ (α), and Q (α) is minimum, and the mathematical model of the problem is expressed as:
min J(h)=(MK(α(h))+MΛ(α(h))+MQ(α(h)))/3 (6)
wherein M isK(α)、MΛ(α)、MQ(α)The number of occurrences of the K (α), Λ (α), Q (α) modes, respectively.
Further, h ═ 0.15 is a good general value for obtaining the high resolution MDVG characteristics, and can be used as a reference value for calculation.
Drawings
FIG. 1 is a visible graph algorithm legend;
FIG. 2 is a process fault monitoring method based on MDVG according to an embodiment of the invention;
FIG. 3 shows the TE process in the embodiment;
FIG. 4 is a relative node mean for a TE process;
FIG. 5 is a relative average connection length for the TE process;
FIG. 6 shows the number of disconnected groups in the TE process;
FIG. 7 is a partial enlarged view of the number of unconnected clusters in the TE process;
FIG. 8 is a process monitoring of DVG-Q (85o) for Fault 3;
FIG. 9 is a process monitoring of MDVG-Q (65o) for Fault 3;
fig. 10 is a process monitoring of LKPCA-T2 for fault 3;
fig. 11 is a process monitoring of LKPCA-Q versus fault 3.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
The process fault monitoring method based on MDVG comprises the following steps:
s101, determining monitoring variables, respectively normalizing the historical data of each variable according to a certain moving pane length, and mapping the variables to a complex network by using an MDVG algorithm.
The conventional process of mapping time series data to a complex network by DVG is generally as follows:
consider any two data (t) in a set of time series datai,xi) And (t)k,xk),i<k, for all data (t) between themj,xj),i<j<k, if NVG visible condition is satisfied
x j < x i + ( x k - x i ) t j - t i t k - t i - - - ( 1 )
And definition of the viewing angle alpha
<math> <mrow> <mi>&alpha;</mi> <mo>&lt;</mo> <msub> <mi>&alpha;</mi> <mi>ik</mi> </msub> <mo>=</mo> <mi>arctan</mi> <mfrac> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> </mrow> <mrow> <msub> <mi>t</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> </math>
Then the two data mapped nodes are considered visible and connected in the network. For example, as shown in fig. 1, a line 1 belongs to all the figures-NVG, DVG (α ═ 90 °), HVG, a line 2 belongs to NVG and HVG but not DVG (α ═ 90 °), and a line 3 belongs to NVG and DVG (α ═ 90 °) but not HVG.
In the process of mapping the time series data to the complex network by the traditional DVG, the influence on the network construction caused by the relation between the time interval and the value size is not considered, so that the network structure difference of the same group of data under different visual angles and different data under the same visual angle is not large, and the further analysis is hindered.
For this purpose, the improved dynamic visibility graph (MDVG) algorithm proposed by the present invention introduces a time interval constant h to further define the viewing angle on the premise of satisfying formula (1), and optimizes formula (2) in the DVG algorithm:
<math> <mrow> <mi>&alpha;</mi> <mrow> <mo>(</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>&lt;</mo> <msub> <mi>&alpha;</mi> <mi>ik</mi> </msub> <mo>=</mo> <mi>arctan</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> </mrow> <mrow> <msub> <mi>t</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> </mrow> </mfrac> <mo>*</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow> </math>
and (3) determining the corresponding unique network structure of the time series data under the viewing angle alpha and the time interval constant h through the calculation of the formulas (1) and (3).
The time interval constant h can be obtained in two ways:
1. the time interval constant h is determined by a particle swarm optimization algorithm, so that the mean value of the times of occurrence of the modes of K (alpha), Lambda (alpha) and Q (alpha) is minimum, and the mathematical model of the problem is expressed as:
min J(h)=(MK(α(h))+MΛ(α(h))+MQ(α(h)))/3 (6)
wherein M isK(α)、MΛ(α)、MQ(α)The number of occurrences of the K (α), Λ (α), Q (α) modes, respectively.
The mode is the value with the largest occurrence frequency in a group of data, and the smaller the occurrence frequency of the mode is, the more the MDVG characteristics are not concentrated on a certain specific value along with the change of the viewing angle, and more abundant information can be obtained from a specific time sequence to perform processing and analysis, so that the mode is easier to distinguish from other time sequences.
2. The average value is obtained through multiple operations on different types of data with different lengths, and found that h is 0.15, which is a good general value capable of obtaining the high-resolution MDVG characteristic and can be used as a reference value for the operations.
S102, calculating three characteristic parameters K (alpha), Lambda (alpha) and Q (alpha) of the mapped network, and determining a monitoring index and a corresponding threshold value.
Three important characteristic parameters characterizing MDVG (α) are as follows:
(1) node relative average degree K (α):
<math> <mrow> <mi>K</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average of the network nodes at the viewing angles alpha and pi, respectively.
(2) Relative average connection length Λ (α):
<math> <mrow> <mi>&Lambda;</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average link lengths of the network at viewing angles alpha and pi, respectively.
(3) Number of non-connected groups Q (α): if a group of nodes arranged in sequence has at least one pair of connection relations between each node and any node in the non-group has no connection relation, the group of nodes forms a non-connected group. The number of the non-connected groups in each network is the number of the non-connected groups.
The monitoring index may include the most value, the average value, or the value at a specific viewing angle of the above three characteristic parameters, and in this step, a threshold value of the monitoring index may be determined, within which the monitoring index is normal.
S103, carrying out online process monitoring, monitoring the current data of each variable by adopting the length of the movable pane which is the same as the historical data, and calculating the monitoring index of the current data.
The monitoring index is consistent with that of step S102, and may include the most value, the average value, or the value at a specific viewing angle of the above three characteristics, and the calculation manner of the three characteristics is as described above.
And S104, judging whether the monitoring index of the current data exceeds a threshold value, and if so, giving an alarm by the system so as to facilitate an operator to search and determine the fault reason in time.
The method of the present application is further described below by way of an example.
The te (tennessee eastman) process model is a realistic simulation program for a family of industrial plants, widely used for standard testing in control and monitoring studies, and is shown in the flow chart of fig. 6. The TE process comprises 5 main units, namely a reactor, a condenser, a compressor, a separator and a stripper, producing 2 products by 4 reactions, while generating a total of 8 components of inert gases and by-products, noted A, B, C, D, E, F, G and H, respectively. The whole system comprises 12 operating variables and 41 process variables (comprising 22 direct measurement variables and 19 analysis variables), and 20 process faults are preset (wherein the types of faults 1-8 are step changes, the faults 9-12 are random changes, the fault 13 is slow drift, the faults 14 and 15 are valve sticking, and the faults 16-20 are unknown). Since the MDVG process monitoring method is based only on process data, it may not be necessary to know the mathematical model of the process in advance.
Selecting 22 direct measurement variables as monitoring variables, enabling the length of a mobile pane to be 50, enabling sampling intervals to be 0.03h, respectively carrying out normalization processing on 1100 sampling data according to the variables, mapping the 1100 sampling data into a complex network with 1100 nodes by utilizing an MDVG algorithm, and then calculating a relevant characteristic value of the network. Four different types of fault data and normal data are selected for analysis and comparison, each type takes 5 groups of data in different panes to reduce randomness, and the calculation results are shown in fig. 3, 4 and 5.
As can be seen from fig. 3, 4, and 5, lines of the same color are grouped together, and lines of different colors are distinguished to some extent, which indicates that the MDVG algorithm can better distinguish different types of data, and the characteristic difference between the same type of data is basically small.
As shown in fig. 6, the MDVG number of unconnected groups Q (α -65 °) at a viewing angle α of 65 ° can be selected as a monitoring index to perform on-line process monitoring on the TE process, where the upper and lower limits of the monitoring threshold are 540 and 520, respectively, that is, when the Q (α -65 °) value of the data is between 520 and 540, the process is considered to be in a normal state, and otherwise the process fails.
The failure 3 is a random change caused by the temperature D in the flow 2, belongs to a failure which is difficult to detect, has a certain representativeness, and is analyzed in detail below. 300 sets of data are collected, a fault 3 is introduced from the 101 th set, and online monitoring results of the DVG and MDVG methods and a local kernel principal component analysis (LKCAS) method which is relatively advanced at present are compared. The results of the operation of each monitoring method are shown in table 1.
TABLE 1 results of the various monitoring methods
According to the simulation result of the TE fault process, the MDVG method can improve the monitoring effect compared with the DVG method under the respective selected relatively optimal monitoring indexes. At the same time, T compared to LKCAA2And Q statistic, the MDVG process monitoring method provided by the invention has higher accuracy and can monitor the occurrence of the fault earlier.
The description of the present invention has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to practitioners skilled in this art. The embodiment was chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.

Claims (4)

1. A method for monitoring process faults based on an improved dynamic visibility diagram, the method comprising the steps of:
s101, determining monitoring variables, respectively normalizing historical data of the variables according to a certain moving pane length, and mapping the variables to a complex network by using an MDVG algorithm;
the process of mapping to a complex network using the MDVG algorithm is as follows:
consider any two data (t) in a set of time series datai,xi) And (t)k,xk),i<k,For all data (t) between themj,xj),i<j<k, if the visible condition is satisfied
x j < x i + ( x k - x i ) t j - t i t k - t i - - - ( 1 )
<math> <mrow> <mi>&alpha;</mi> <mrow> <mo>(</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>&lt;</mo> <msub> <mi>&alpha;</mi> <mi>ik</mi> </msub> <mo>=</mo> <mi>arctan</mi> <mrow> <mo>(</mo> <mfrac> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> </mrow> <mrow> <msub> <mi>t</mi> <mi>k</mi> </msub> <mo>-</mo> <msub> <mi>t</mi> <mi>i</mi> </msub> </mrow> </mfrac> <mo>*</mo> <mi>h</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow> </math>
Then, the nodes to which the two data maps are considered to be visible and connected in the network,
and (3) determining the corresponding unique network structure of the time series data under the viewing angle alpha and the time interval constant h through the calculation of the formulas (1) and (3).
S102, calculating three characteristic parameters K (alpha), Lambda (alpha) and Q (alpha) of the mapped network, and determining a monitoring index and a corresponding threshold value.
Three important characteristic parameters characterizing MDVG (α) are as follows:
(1) node relative average degree K (α):
<math> <mrow> <mi>K</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>k</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average of the network nodes at the viewing angles alpha and pi, respectively.
(2) Relative average connection length Λ (α):
<math> <mrow> <mi>&Lambda;</mi> <mrow> <mo>(</mo> <mi>&alpha;</mi> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&alpha;</mi> </msub> <mo>/</mo> <msub> <mover> <mi>l</mi> <mo>&OverBar;</mo> </mover> <mi>&pi;</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow> </math>
wherein,andthe average link lengths of the network at viewing angles alpha and pi, respectively.
(3) Number of non-connected groups Q (α): if a group of nodes arranged in sequence has at least one pair of connection relations between each node and any node in the non-group has no connection relation, the group of nodes form a non-connected group, and the number of the non-connected groups in each network is the number of the non-connected groups.
S103, carrying out online process monitoring, monitoring the current data of each variable by adopting the length of a movable pane which is the same as the historical data, and calculating the monitoring index of the current data;
and S104, judging whether the monitoring index of the current data exceeds a threshold value, and if so, giving an alarm by the system so as to facilitate an operator to search and determine the fault reason in time.
2. The method of claim 1, wherein the time interval constant h can be determined by a particle swarm optimization algorithm such that the mean of the number of occurrences of the modes of K (α), Λ (α), Q (α) is minimized, and the mathematical model of the problem is represented as:
minJ(h)=(MK(α(h))+MΛ(α(h))+MQ(α(h)))/3 (6)
wherein M isK(α)、MΛ(α)、MQ(α)The number of occurrences of the K (α), Λ (α), Q (α) modes, respectively.
3. The method of claim 1, wherein h-0.15 is a good general value for obtaining the high resolution MDVG characteristic, and can be used as a reference value for the operation.
4. The method of claim 1, wherein the monitoring index of step S103 is consistent with the monitoring index of step S102, and may include the most value, the average value or the value at a specific viewing angle of the three characteristic parameters.
CN201410806907.7A 2014-12-22 2014-12-22 Process fault monitoring method based on improved dynamic visible graph Active CN104503441B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410806907.7A CN104503441B (en) 2014-12-22 2014-12-22 Process fault monitoring method based on improved dynamic visible graph

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410806907.7A CN104503441B (en) 2014-12-22 2014-12-22 Process fault monitoring method based on improved dynamic visible graph

Publications (2)

Publication Number Publication Date
CN104503441A true CN104503441A (en) 2015-04-08
CN104503441B CN104503441B (en) 2017-02-22

Family

ID=52944845

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410806907.7A Active CN104503441B (en) 2014-12-22 2014-12-22 Process fault monitoring method based on improved dynamic visible graph

Country Status (1)

Country Link
CN (1) CN104503441B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110472335A (en) * 2019-08-15 2019-11-19 中国科学院工程热物理研究所 Sensor fault diagnosis Threshold based on particle swarm optimization algorithm

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002082713A (en) * 2000-07-15 2002-03-22 Intevep Sa Method and system for monitoring industrial process
US6725181B2 (en) * 2001-06-08 2004-04-20 Hon Hai Precision Ind. Co., Ltd. Method and system for collecting and monitoring shop floor information
US20040249491A1 (en) * 2001-08-06 2004-12-09 Hott Brandt Powell Internet wide distributed data control system
CN101446827A (en) * 2008-11-06 2009-06-03 西安交通大学 Process fault analysis device of process industry system and method therefor
JP2010170342A (en) * 2009-01-22 2010-08-05 Mitsubishi Electric Corp Fault monitoring apparatus and method
CN101964021A (en) * 2010-09-29 2011-02-02 东北大学 Penicillin fermentation process failure monitoring method based on recursive kernel principal component analysis
CN103853152A (en) * 2014-03-21 2014-06-11 北京工业大学 Batch process failure monitoring method based on AR-PCA (Autoregressive Principal Component Analysis)
CN103970092A (en) * 2014-04-13 2014-08-06 北京工业大学 Multi-stage fermentation process fault monitoring method based on self-adaption FCM algorithm
CN104133991A (en) * 2014-07-15 2014-11-05 东北大学 Penicillin fermentation process fault diagnosis method based on kernel partial least squares reconstitution

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002082713A (en) * 2000-07-15 2002-03-22 Intevep Sa Method and system for monitoring industrial process
US6725181B2 (en) * 2001-06-08 2004-04-20 Hon Hai Precision Ind. Co., Ltd. Method and system for collecting and monitoring shop floor information
US20040249491A1 (en) * 2001-08-06 2004-12-09 Hott Brandt Powell Internet wide distributed data control system
CN101446827A (en) * 2008-11-06 2009-06-03 西安交通大学 Process fault analysis device of process industry system and method therefor
JP2010170342A (en) * 2009-01-22 2010-08-05 Mitsubishi Electric Corp Fault monitoring apparatus and method
CN101964021A (en) * 2010-09-29 2011-02-02 东北大学 Penicillin fermentation process failure monitoring method based on recursive kernel principal component analysis
CN103853152A (en) * 2014-03-21 2014-06-11 北京工业大学 Batch process failure monitoring method based on AR-PCA (Autoregressive Principal Component Analysis)
CN103970092A (en) * 2014-04-13 2014-08-06 北京工业大学 Multi-stage fermentation process fault monitoring method based on self-adaption FCM algorithm
CN104133991A (en) * 2014-07-15 2014-11-05 东北大学 Penicillin fermentation process fault diagnosis method based on kernel partial least squares reconstitution

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
I.V.BEZSUDNOV等: "From time series to complex networks: the Dynamical Visibility Graph", 《ARXIV PREPRINT ARXIV》 *
ZHONGKE GAO等: "Visibility Graph Analysis of Fluid Flow Signals", 《2012 IEEE FIFTH INTERNATIONAL CONFERENCE ON ADVANCE COMPUTATIONAL INTELLIGENCE》 *
唐镇等: "多重分形序列的可见图分析", 《上海理工学报》 *
陈雨等: "基于数据复杂网络理论的系统故障检测方法", 《化工学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110472335A (en) * 2019-08-15 2019-11-19 中国科学院工程热物理研究所 Sensor fault diagnosis Threshold based on particle swarm optimization algorithm
CN110472335B (en) * 2019-08-15 2022-10-21 中国科学院工程热物理研究所 Sensor fault diagnosis threshold value determination method based on particle swarm optimization algorithm

Also Published As

Publication number Publication date
CN104503441B (en) 2017-02-22

Similar Documents

Publication Publication Date Title
CN111596604B (en) Intelligent fault diagnosis and self-healing control system and method for engineering equipment based on digital twinning
CN116129366B (en) Digital twinning-based park monitoring method and related device
CN106872657B (en) A kind of multivariable water quality parameter time series data accident detection method
KR101316486B1 (en) Error detection method and system
CN107561997B (en) A kind of power equipment state monitoring method based on big data decision tree
Mounce et al. Pattern matching and associative artificial neural networks for water distribution system time series data analysis
CN107273924B (en) Multi-data fusion power plant fault diagnosis method based on fuzzy clustering analysis
CN109858140B (en) Fault diagnosis method for water chilling unit based on information entropy discrete Bayesian network
CN101477374B (en) Continuous casting bleed-out time sequence spacing combined diagnosis prediction method based on fuzzy neural network
CN102282552A (en) System, method and computer program for pattern based intelligent control, monitoring and automation
CN113919540B (en) Method for monitoring operation state of production process and related equipment
Wang et al. Fault detection and identification using a Kullback-Leibler divergence based multi-block principal component analysis and Bayesian inference
CN107257351A (en) One kind is based on grey LOF Traffic anomaly detections system and its detection method
CN117421582A (en) Equipment health analysis method based on multi-source data driving
WO2019071438A1 (en) Method and apparatus for monitoring state of device in process industry and medium
JP2015109028A (en) Data relevancy analysis system and method
CN117560300B (en) Intelligent internet of things flow prediction and optimization system
CN104317778A (en) Massive monitoring data based substation equipment fault diagnosis method
CN111126477A (en) Learning and reasoning method of hybrid Bayesian network
Carratù et al. A novel methodology for unsupervised anomaly detection in industrial electrical systems
CN104503441B (en) Process fault monitoring method based on improved dynamic visible graph
CN117056533A (en) Knowledge graph-based industrial knowledge mining modeling method and system
Tolas et al. Periodicity detection algorithm and applications on IoT data
Wang et al. Multi‐block principal component analysis based on variable weight information and its application to multivariate process monitoring
CN115310499B (en) Industrial equipment fault diagnosis system and method based on data fusion

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant