CN105404709A - Complex network based sensitive monitoring point analysis method for dam health monitoring - Google Patents
Complex network based sensitive monitoring point analysis method for dam health monitoring Download PDFInfo
- Publication number
- CN105404709A CN105404709A CN201510691360.5A CN201510691360A CN105404709A CN 105404709 A CN105404709 A CN 105404709A CN 201510691360 A CN201510691360 A CN 201510691360A CN 105404709 A CN105404709 A CN 105404709A
- Authority
- CN
- China
- Prior art keywords
- measuring point
- complex network
- measured value
- dykes
- dams
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Theoretical Computer Science (AREA)
- Civil Engineering (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Architecture (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a complex network based sensitive monitoring point analysis method for dam health monitoring. The method comprises the steps of obtaining a monitoring value time sequence reflecting a health condition of engineering through actual monitoring data corresponding to each monitoring item according to dam health monitoring items; converting the monitoring value time sequence into a symbol sequence with an equiprobable coarse-graining method; coding the symbol sequence in a topological structure of a complex network to construct the complex network; and analyzing feature parameters of the complex network to obtain sensitive monitoring points of the dam health monitoring items and a condition evolution law of a dam. The proposed novel method has the capabilities of effectively identifying important monitoring points and judging dam condition evolution, overcomes the sensitivity of a conventional analysis method to data errors, and can discover the importance of each monitoring point in a multi-scale manner, realize analysis on a correlation relationship between the monitoring points, master the health condition of dam engineering timely and accurately and realize more effective early warning of the dam condition.
Description
Technical field
The present invention relates to a kind of dykes and dams health monitoring analytical approach, particularly relate to the responsive measuring point of a kind of dykes and dams health monitoring based on complex network and analyze defining method.
Background technology
Dykes and dams health monitoring is not only conducive to avoiding large dangerous situation and disaster to occur, and is also Feedback Design construction quality and the precondition checking scientific and technical result to the accurate grasp of dykes and dams condition and subtle change.Grasping the condition of dykes and dams in time, exactly, is the basis ensureing that embankment safety runs.
Dykes and dams health monitoring is the knowledge that lake, river and reservoir safety run, but at present, overwhelming majority field data analytical approach is higher to data demand, as required, data meet the requirements such as stationarity, normality and independent same distribution, and actual field data all has non-stationary, correlativity and heteroscedasticity, thus make existing analytical approach be difficult to accurately grasp measured value rule, sometimes even obtain wrong conclusion.Existing analytical approach is determined often helpless to responsive or important measuring point analysis simultaneously, can only rely on expertise.
In fact, the actual measured value sequence of dykes and dams health monitoring is scarcely steady, especially dykes and dams critical days or very operating mode phase.And the determination of responsive or important measuring point often isostructuralism state analysis be mutually related, be two aspects of a problem.Current conventional method of analysis is many all can only adapt to stationary time series or to data noise-sensitive, thus can not identify responsive measuring point, effectively can not analyze structure behaviour.
Summary of the invention
Fundamental purpose of the present invention is, overcome deficiency of the prior art, make full use of Complex Networks Analysis method multiscale analysis function and to the insensitive advantage of error, there is provided a kind of dykes and dams health monitoring based on complex network responsive measuring point analytical approach, overcome the susceptibility of conventional method of analysis to data error, the significance level of each measuring point can be found, realize the analysis of correlationship between each measuring point, and the healthy condition of the engineering grasping dykes and dams exactly, realize effective early warning of dykes and dams condition.
In order to achieve the above object, the technical solution adopted in the present invention is:
The responsive measuring point analytical approach of dykes and dams health monitoring based on complex network, comprises the following steps:
1) according to dykes and dams health monitoring project and measuring point, application is laid in the actual measured value time series of the corresponding measuring point of sensor acquisition dykes and dams of measuring point;
2) use the method for equiprobability coarse, measured value time series is changed into symbol sebolic addressing;
3) symbol sebolic addressing is coded in the topological structure of complex network, complex structure networking;
4) analyze the characteristic parameter of complex network, obtain responsive measuring point and the condition Evolution thereof of this dykes and dams health monitoring project, the healthy condition of engineering of carrying out dykes and dams differentiates.
The present invention is set to further: described step 1) in dykes and dams health monitoring project comprise environment parameter, seepage field, temperature field and deformation field.
The present invention is set to further: described step 2) method of equal probability coarse, specifically comprise step, 2-1) establish x
maxand x
minbe measured value seasonal effect in time series maximal value and minimum value respectively, the symbol after conversion has N kind, is designated as s
1, s
2..., s
n, definition
S
i＝s
j，x
min+(j-1)d≤x
i＜x
min+j
d，j＝1，...，N(1)
Wherein, N>2, d=(x
max-x
min)/N; The symbol sebolic addressing that N kind kinds of characters forms is obtained by formula (1);
2-2) remember dykes and dams measured value time series for x (t)=1 ..., N}, through type (2) calculates measured value time series fluctuation k (t),
Wherein, Δ t is the time interval;
2-3) through type (3) calculates the Probability p (k) that different undulating quantity may occur,
Wherein, the number of times of the fluctuation mode x generation of the corresponding measured value sequence of Num (x);
2-4) measured value seasonal effect in time series fluctuation k (t) is divided into 5 intervals, defining 5 characteristic characters is formula (4),
Wherein, t represents measured value numerical value to be increased fast, and r represents measured value numerical value to be increased slowly, and e represents measured value numerical value and do not increase and do not subtract, and d represents measured value numerical value to be reduced slowly, and f represents measured value numerical value to be reduced fast;
Thus measured value time series to be converted to symbol sebolic addressing be formula (5),
S＝(S
1S
2S
3...)，S
i∈(t，r，e，d，f)(5)
The present invention is set to further: described step 3) middle complex structure networking, be after the topological structure of complex network symbol sebolic addressing being coded in oriented weighting, convert the accessible form of pajek software to by excle2pajek software.
The present invention is set to further: described step 4) in the characteristic parameter of complex network is analyzed, undertaken by pajek software.
The present invention is set to further: described step 4) in the characteristic parameter of complex network comprise average path length, on average gather coefficient and modularity coefficient.
The present invention is set to further: described environment parameter comprises level of tail water measuring point and atmospheric pressure measuring point before upper pond level measuring point before dam break, dam break; Described seepage field comprises osmotic pressure water level measuring point, seepage flow measuring point and dam body moisture measuring point; Described temperature field adopts distribution type fiber-optic to measure, and comprises dam body surface temperature measuring point, storehouse water temperature measuring point, dam body internal temperature measuring point, upstream water temperature measuring point and downstream water temperature measuring point; Described deformation field comprises dam body surface deformation measuring point, dam body internal modification measuring point and inclination measuring point.
Compared with prior art, the beneficial effect that the present invention has is:
According to dam-break experiments measured data; the dynamic characteristic of measured value time series variation is disclosed from the angle of complex network; first the method for equiprobability coarse is used; measured value time series is changed into symbol sebolic addressing; symbol sequence is mapped to network parameter complex structure networking, then by the analysis to the characteristic parameter of complex network, obtains important measuring point and condition Evolution thereof; thus realize the healthy singular analysis of dykes and dams, have and effectively identify important measuring point and differentiate the ability that dykes and dams condition develops.Wherein, adopt coarse method, be conducive to overcoming the impact of data error on analysis result, be conducive to the analysis realizing different scale; Adopt Complex Networks Analysis method simultaneously, the significance level of each measuring point can be found, realize the analysis of correlationship between each measuring point, thus extract responsive measuring point, grasp the healthy condition of engineering of dykes and dams timely and accurately, realize more effective early warning of dykes and dams condition.
Foregoing is only the general introduction of technical solution of the present invention, and in order to clearer understanding technological means of the present invention, below in conjunction with accompanying drawing, the invention will be further described.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the responsive measuring point analytical approach of dykes and dams health monitoring that the present invention is based on complex network;
Fig. 2 is the environment temperature graph of retaining phase in the dykes and dams process of the test of the embodiment of the present invention;
Fig. 3 is the stage hydrograph of process of bursting in the dykes and dams process of the test of the embodiment of the present invention;
Fig. 4 is that the dam break of the embodiment of the present invention is front upper, the complex network figure of level of tail water measured value;
Fig. 5 is the complex network figure that whole process seepage pressure measuring point M1, M2 of the embodiment of the present invention is corresponding;
Fig. 6 is the complex network figure that dam break phase seepage pressure measuring point M1, M2 of the embodiment of the present invention is corresponding;
Fig. 7 is the complex network figure of the fiber optic temperature measured value of 4 different measuring points of the embodiment of the present invention;
Fig. 8 is the complex network figure that dam break phase measuring point S1, S2 measured value of the embodiment of the present invention is corresponding;
Fig. 9 is the complex network figure that whole process inclination measuring point K1, K2 measured value of the embodiment of the present invention is corresponding;
Figure 10 is the complex network figure that dam break phase inclination measuring point K1, K2 measured value of the embodiment of the present invention is corresponding;
Figure 11 is the complex network figure that three groups of temperature measured values of the whole process of the embodiment of the present invention are corresponding.
Embodiment
Below in conjunction with Figure of description, the present invention is further illustrated.
As shown in Figure 1, the invention provides the responsive measuring point analytical approach of a kind of dykes and dams health monitoring based on complex network, comprise the following steps:
1) the project application monitoring sensor set by dykes and dams health monitoring, obtains the measured value time series of the measuring point of this dykes and dams health monitoring project; Wherein, health monitoring project in dam comprises environment parameter, seepage field, temperature field and deformation field etc. as required.
2) use the method for equiprobability coarse, measured value time series is changed into symbol sebolic addressing; Wherein, the method for equiprobability coarse, specifically comprises step,
2-1) establish x
maxand x
minbe measured value seasonal effect in time series maximal value and minimum value respectively, the symbol after conversion has N kind, is designated as s
1, s
2..., s
n, definition
S
i＝s
j，x
min+(j-1)d≤x
i＜x
min+jd，j＝1，...，N(1)
Wherein, N>2, d=(x
max-x
min)/N; The symbol sebolic addressing that N kind kinds of characters forms is obtained by formula (1);
2-2) remember dykes and dams measured value time series for x (t)=1 ..., N}, through type (2) calculates measured value time series fluctuation k (t),
Wherein, Δ t is the time interval;
2-3) through type (3) calculates the Probability p (k) that different undulating quantity may occur,
Wherein, the number of times of the fluctuation mode x generation of the corresponding measured value sequence of Num (x);
2-4) measured value seasonal effect in time series fluctuation k (t) is divided into 5 intervals, defining 5 characteristic characters is formula (4),
Wherein, t represents measured value numerical value to be increased fast, and r represents measured value numerical value to be increased slowly, and e represents measured value numerical value and do not increase and do not subtract, and d represents measured value numerical value to be reduced slowly, and f represents measured value numerical value to be reduced fast;
Thus measured value time series to be converted to symbol sebolic addressing be formula (5),
S＝(S
1S
2S
3...)，S
i∈(t，r，e，d，f)(5)
3) symbol sebolic addressing is coded in the topological structure of complex network of oriented weighting, converts the accessible form of pajek software to by excle2pajek software afterwards, complex structure networking.
4) analyzed by the characteristic parameter of pajek software to complex network, obtain responsive measuring point and the condition Evolution thereof of this dykes and dams health monitoring project, the healthy condition of engineering of carrying out dykes and dams differentiates.
Wherein, the characteristic parameter of complex network comprises average path length, on average gathers coefficient and modularity coefficient.Average path length represents a kind of mode to the time required for another kind of MODAL TRANSFORMATION OF A, and therefore shorter average path length can show the correlativity of short distance; On average gather coefficient and then illustrate the speed changed between fluctuation model, the larger coefficient that on average gathers represents that conversion frequently; The simultaneously sparse degree that contacts each other of the numerical representation method network internal group variety of modularity, larger modularity represents to have stronger corporations' relation.
Each N word string is considered as a summit of network, between two N word strings adjacent in symbol sebolic addressing, there is a limit, point to a rear N word string by previous N word string.If there are many limits between two summit i and j, so limit number is exactly the weights W on the limit between these two summits
ij.Suppose that network divide into m corporations, the matrix c on structure m × m rank, c
ijrepresent the ratio that the limit connecting Liang Ge corporations i and corporations j is occupied in complex network, then modularity can be expressed as:
Wherein, the value of 0≤M < 1, M, more close to 1, means that partitioning algorithm is better.
The embodiment of the present invention:
Test dam maximum storage capacity reaches 10 × 10
4m
3, barrage is viscosity homogeneous earth dam, and maximum height of dam reaches 9.7m, the long 120m of dam crest, wide 3m, and spillway is positioned on the right side of dam, wide 5m, and there is the culvert that discharges water of aperture 0.3m in left side.
Dykes and dams health monitoring project comprises environment parameter, seepage field, temperature field and deformation field, environment parameter comprises level of tail water measuring point and atmospheric pressure measuring point before upper pond level measuring point before dam break, dam break, seepage field comprises osmotic pressure water level measuring point, seepage flow measuring point and dam body moisture measuring point, temperature field comprises dam body surface temperature measuring point, storehouse water temperature measuring point, dam body internal temperature measuring point, upstream water temperature measuring point and downstream water temperature measuring point, and deformation field comprises dam body surface deformation measuring point, dam body internal modification measuring point and inclination measuring point.
The process of the test of test dykes and dams comprises dam filing phase, the first water impoundment phase, stable level phase, the second time process such as retaining phase and the phase of bursting, engineering lasts 46 days, it is as shown in table 1 that process of the test lasts statistical form, in process of the test, as shown in Figure 2, the stage hydrograph of process of bursting in process of the test as shown in Figure 3 for the environment temperature graph of retaining phase.
Table 1
Wherein, carry out instrument embedding in the dam filing phase, lay level of tail water measuring point 1, dam body surface temperature measuring point 1, storehouse water temperature measuring point 1, dam body internal temperature measuring point 1, upstream water temperature measuring point 1, downstream water temperature measuring point 1, atmospheric pressure measuring point 1, osmotic pressure water level measuring point 13, seepage flow measuring point 1, dam body moisture measuring point 2, dam body surface deformation measuring point 6, dam body internal modification measuring point 9 before upper pond level measuring point 1 before dam break, dam break altogether; And temperature field adopts distribution type fiber-optic to measure, spatial resolution 1m, the temperature survey line 20m of dam body surface temperature measuring point, the temperature survey line 20m of storehouse water temperature measuring point, the temperature survey line 343.1m of dam body internal temperature measuring point.
According to dykes and dams health monitoring project, its measured value time series is obtained respectively in the process of the test of test dykes and dams, convert measured value time series to symbol sebolic addressing, as in the dam of 2012/10/613:00:08 to 2012/10/620:00:07, the measured data of osmotic pressure M1 measuring point is 43.25, 43.25, 43.25, 43.25, 43.25, 43.26, 43.26, 43.26, 43.26, 43.26, 43.28, 43.28, 43.28, 43.31, 43.3, 43.29, 43.29, 43.29, 43.22, 43.22, 43.22, 43.22, 43.22, 43.13, 43.13, 43.13, 43.13, 43.13, 43.14, the symbol sebolic addressing then converted to is: feeeereeeeteetddeefeeeefeeeer.
Carry out complex network structure and analysis afterwards, obtain the responsive measuring point in different measuring points and condition Evolution thereof, the healthy condition of engineering of carrying out dykes and dams differentiates; The following measuring point according to different monitoring project describes in detail respectively.
1, the analysis of level of tail water measuring point before upper pond level measuring point and dam break before dam break
Data before and after dam break are separately built complex network, show that dam break is front upper, the complex network figure of level of tail water Monitoring Data, as shown in Figure 4, in Fig. 4, (a) is the complex network figure of upper pond level measured value before dam break, and in Fig. 4, (b) is the complex network figure of level of tail water measured value before dam break.
In the complex network that upper pond level measured value as shown in Fig. 4-(a) is corresponding, thicker lines have: ee → ee, ee → re, er → ee, re → ee, these change procedures account for 69.3% of sum, and on the time, these changes appear at October 16 to November 16, and the actual conditions risen with retaining phase water level conform to.In the complex network of the level of tail water measured value as shown in Fig. 4-(b), the limit that thicker line is corresponding is: ee → ee, re → ee, ee → er, ee → re, these changes account for 48.6% of change sum, and on the time, these changes appear at the retaining phase of October 14 to November 14.
Find that the frequency that r occurs is all higher by word statistics frequently.As can be seen from above statistics, upper pond level data comprise more information than level of tail water data.
Two groups of back end degree are added up, the dam break be shown in Table 2 are front upper, level of tail water measured value node degree statistical form.
Table 2
In his-and-hers watches 2, these nodes carry out statistics and find, main 6 node degrees of upper pond level account for 98.85% of sum, and main 6 node degrees of the level of tail water account for 92.58% of node total number, and visible upper pond level measured value is more frequent in certain several intervals change.
Dam break phase upper pond level constantly reduces, the level of tail water then moment dam break phase water level rise to some extent.Because dam break phase waterlevel data is relatively less, be difficult to be depicted as complex network figure.The topological property of data before the water level dam break of upstream and downstream is analyzed, the complex network index of data before upstream and downstream water level dam break as shown in table 3 below.
Table 3
Can be found out by table 3, the complex network that upper pond level measured value is corresponding is all greater than downstream measuring point in three, and visible upper pond level measuring point comprises more information, so upper pond level measuring point is typical position before dam break, tackles it and carries out emphasis monitoring.
2, the analysis of seepage pressure
Osmotic pressure water level measuring point M1, M2 are shown in Fig. 5 at the complex network that whole dam-break experiments process measured value is corresponding, in Fig. 5, (a) is the complex network figure of whole process osmotic pressure water level measuring point M1 measured value, and in Fig. 5, (b) is the complex network figure of whole process osmotic pressure water level measuring point M2 measured value.
First analyze node degree, the degree of a node means that more greatly this node is just even more important in some sense.Table 4, table 5 are enumerated the sequence of the node degree size of M1, M2 respectively.
Table 4
Table 5
As can be seen from table 4 and table 5, node er, re, ed, the node number of degrees of rd, de, dr are larger, this illustrates that the fluctuation mode representated by these nodes serves important correlation in dykes and dams condition change procedure, and various fluctuation mode changes high by the frequency of this several fluctuation MODAL TRANSFORMATION OF A in other words to this several fluctuation mode.Carry out arrangement to the time period that this several mode occurs respectively in addition to find, the change mode of M1 focused mostly on November 13 to November 17, latter stage of namely second time retaining and burst during the phase, obvious several change has: re → ee, ee → er, ee → rd, these changes have accounted for 21.7% of change sum, and the change mode of M2 focuses mostly in November 15 to November 17, more corresponding than the former will rearward, significantly change has: ee → er, re → ee, de → ee, these changes account for 20.3% of change sum.Can be found by the change mode of node, the change of M1 has more early more concentrated character, but statistics and the M1 of M2 node are also more or less the same, illustrate that M1, M2 have identical feature in the change of reaction dykes and dams condition, but early stage, M1 was more responsive as dam structure abnormity early warning.
Osmotic pressure water level measuring point M1, M2 are shown in Fig. 6 at the complex network that dam break phase measured value is corresponding, in Fig. 6, (a) is the complex network figure of dam break phase osmotic pressure water level measuring point M1 measured value, and in Fig. 6, (b) is the complex network figure of dam break phase osmotic pressure water level measuring point M2 measured value.
The dam break phase complex network Indexes Comparison corresponding with whole process is in table 6, and table 6 is seepage pressure data topology character Complex Networks Feature value contrast table.
Table 6
As can be seen from Table 6, the complex network of M2 has and larger on average gathers coefficient and less average path length.Average path length illustrated by a kind of mode to the time required for another kind of MODAL TRANSFORMATION OF A, this value is close to 1, illustrate that the switching time of this pattern is approximately about 2 days, show the correlativity in order to short distance, this has certain reference value to the short-term forecasting of the healthy measured value of dykes and dams.The larger small-scale group variety on average gathering coefficient and then show to exist between various fluctuation mode, and the key of group variety inside is better, and the patten transformation showing as fluctuation is comparatively frequent.Trying to achieve this mixed-media network modules mixed-media degree according to formula (6) is 0.10476, and close to 0, therefore the group variety relation of this network internal shows comparatively sparse, thinks to have more weak community structure.And the data of M2 measuring point dam break phase have shorter average path length and larger on average gather coefficient, and the data of M1 measuring point of comparing then do not have this representativeness.According to the analysis to this several topological property, reflect the frequent of M2 measuring point measured value sequence fluctuation, the conversion that what the reacting condition of these data went out is exactly between various fluctuation mode, better can hold the variation of measured value sequence, thus the dam break phase should carry out emphasis monitoring to M2 measuring point.
3, the analysis of temperature in dam
Be illustrated in figure 7 the complex network figure of the fiber optic temperature measured value of 4 different measuring points, in Fig. 7, (a) is complex network figure corresponding to TF-1 measuring point, in Fig. 7, (b) is complex network figure corresponding to TF-2 measuring point, in Fig. 7, (c) is complex network figure corresponding to TF-3 measuring point, and in Fig. 7, (d) is complex network figure corresponding to TF-4 measuring point.
Find that measuring point TF1 occupies better proportion in these typical Mode variations by the comparison of complex network corresponding to 4 fiber optic temperature measured values, for: 41.5%.Rd → dr is found in addition according to statistics, dr → rd, rd → rr, rr → rd, the frequencies that 4 kinds of changes occur are the highest, be embodied in complex network figure be exactly corresponding lines i.e. complex network limit more slightly, and measuring point TF-1 occupies 40.22% respectively in these 4 kinds of versions, 40.48%, 42.59%, 42.59%.Complex network figure shown in Fig. 7 is constructed respectively to 4 measuring points, contrasts with conceptual data after then carrying out the analysis of topological property, the fiber optic temperature be shown in Table 7 each measuring point measured value complex network Indexes Comparison.
| TF-1 | TF-2 | TF-3 | TF-4 | |
| Average path length | 1.00708 | 1.04545 | 1.04444 | 1.09091 |
| On average gather coefficient | 0.98291 | 0.955222 | 0.96111 | 0.92015 |
| Modularity coefficient | 0.00193 | 0.00165 | 0.00265 | 0.00638 |
Table 7
Can be found by Fig. 7 and table 7, measuring point TF-1 changes the point concentrated the most, and therefore measuring point TF-1 can as the typical position of optical fiber temperature-measurement.Carry out brief analysis to the topological property of fiber optic temperature data below, the average path length of measuring point TF-1 data is 1.00708, and on average gathering coefficient is 0.98291.As can be seen here, the complex network of measuring point TF-1 has and larger on average gathers coefficient and less average path length.Average path length illustrated by a kind of mode to the time required for another kind of MODAL TRANSFORMATION OF A, this value is close to 1, illustrate that the switching time of this pattern is approximately about 2 days, show the correlativity in order to short distance, the larger coefficient that on average gathers then has showed the frequent of data fluctuations.Trying to achieve this mixed-media network modules mixed-media degree in combined data according to formula (6) is 0.00168, and have relatively very little modularity, therefore think there is not obvious community structure between 4 measuring points, group variety sex expression must be more sparse.Thus according to the analysis of topological property, show that TF-1 is typical position, advise the monitoring carrying out emphasis during dykes and dams monitoring.
4, the analysis of water percentage
Be illustrated in figure 8 complex network figure corresponding to dam break phase measuring point S1, S2, in Fig. 8, (a) is complex network figure corresponding to dam break phase measuring point S1, and in Fig. 8, (b) is complex network figure corresponding to dam break phase measuring point S2.The sequence of node degree is carried out as shown in table 8 and table 9 to the node shown in Fig. 8.
Table 8
Table 9
As can be seen from table 8 and table 9, node rr, er, the node degree of the nodes such as re, de, rd is larger, illustrate that the fluctuation mode of these node on behalf serves important correlation among a series of data movement of the dam break on simulation dam, various mode is high to the frequency of this several MODAL TRANSFORMATION OF A.Can find out that main version has: ee → ee intuitively by image, rr → rr, ee → er, re → ee, these changes account for 44.5% of change sum, analyze rear discovery in addition to the time period that these change mode occur, the change mode wherein representing S1 in unconverted ee → ee focused mostly on October 15 to November 16, corresponding is exactly the retaining phase and bursts the phase, and the S2 that compares focuses mostly on October 14 to November 16, little with S1 gap.And rr → rr version that representative is risen continuously slowly, S1 then focuses mostly in October 30 to November 10, and S2 then concentrates on November 5 to November 13, is not also clearly in macroscopically difference.S1 and S2 two groups of data are depicted as complex network Fig. 8 respectively analyze, and contrast the topological property of measuring point S1, S2 dam break issue certificate as shown in table 10 and comparing of conceptual data.
| Complex network index | Overall | S1 | S2 |
| Average path length | 1.16667 | 1.03636 | 1.10256 |
| On average gather coefficient | 0.95411 | 0.96522 | 0.90857 |
| Modularity coefficient | 0.16779 | 0.00167 | 0.00631 |
Table 10
Can be found out by obvious, S1 complex network has and larger on average gathers coefficient and less average path length.Change frequently between this network mode as can be seen here, show as the correlativity of short distance, there is small-scale group variety between various fluctuation mode in network, and the internal correlation of group variety is better, the patten transformation showing as fluctuation is more frequent simultaneously.The modularity of the network simultaneously asked by formula (6) is 0.00167, and by comparing discovery, this network has very little modularity, and therefore inner group variety Relationship Comparison is sparse, has more weak community structure.According to the analysis of several topological property of appeal, the conversion drawing the change that measuring point S1 better can react measured value sequence and fluctuate between mode, therefore measuring point S1 is typical position, should carry out emphasis monitoring in daily monitoring.
5, the analysis of inclination
Inclination measuring point K1, K2 are at complex network figure corresponding to whole process measured value as shown in Figure 9, in Fig. 9, (a) is complex network figure corresponding to whole process inclination measuring point K1, and in Fig. 9, (b) is complex network figure corresponding to whole process inclination measuring point K2.Node degree shown in Fig. 9 is added up, as shown in table 11 and table 12.
Table 11
Table 12
Can be seen by table 11 and table 12, between inclination measuring point K1, K2, node degree difference is larger, and the node degree statistics of K1 is concentrated relatively more, and the statistics of the node degree of K2 is want comparing dispersion.Obviously can be obtained significantly being changed in inclination data K1 by complex network figure: ee → ee, re → ee, er → ee, ee → re, these changes have accounted for 38.0% of change sum, times that these changes occur are carried out to statistics discovery and focus mostly on November 13 to November 17, latter stage of namely second time retaining with burst the phase.And significantly severally in the change mode of K2 to be changed to: ft → rd, rf → tr, rf → tf, these changes only account for 1.2% of change sum.Therefore tentatively judging that the change mode of K1 is more concentrated, is important measuring point.
The complex network figure that inclination measuring point K1, K2 obtain in the measured value data of dam break phase as shown in Figure 10, in Figure 10, (a) is the complex network figure of dam break phase inclination measuring point K1 measured value, and in Figure 10, (b) is the complex network figure of dam break phase inclination measuring point K2 measured value.The dam break phase complex network Indexes Comparison corresponding with whole process is in table 13, and table 13 is inclination data topology character Complex Networks Feature value contrast table.
| Topological property/measuring point and period | The whole process of K1 | The whole process of K2 | The K1 dam break phase | The K2 dam break phase |
| Average path length | 1.04070 | 1.12944 | 1.28788 | 1.35498 |
| On average gather coefficient | 0.97138 | 0.92783 | 0.83836 | 0.76318 |
| Modularity coefficient | 0.00150 | 0.00773 | 0.12245 | 0.03991 |
Table 13
Can clearly be seen that complex network that K1 measuring point data is formed has by table 13 and larger on average gather coefficient and less average path length.Average path length value is close to 1, illustrate that the switching time of this pattern is approximately about 2 days, show the correlativity in order to short distance, the larger coefficient that on average gathers then illustrates and there are relation some group varietys comparatively closely between various fluctuation mode, and the patten transformation between them is more frequent.Trying to achieve this mixed-media network modules mixed-media degree according to formula (6) is 0.00150, close to 0, therefore thinks not have community structure by the comparatively sparse of each group variety of this network internal relation performance each other.According to above-mentioned complex network index, K1 measuring point is typical position, should carry out the monitoring of emphasis in daily monitoring.
6, the analysis of upstream water temperature, downstream water temperature and temperature
Complex network corresponding to three groups of temperature measured value time serieses of whole process of the test as shown in figure 11, in Figure 11, (a) is the complex network figure of upstream water temperature measured value, in Figure 11, (b) is the complex network figure of downstream water temperature measured value, the complex network figure that in Figure 11, (c) is temperature measured value.
The change that lines thicker in the complex network that upstream water temperature measured value is corresponding are corresponding has: ee → ee, de → ee, ee → ed, ee → de, these changes account for 72.0% of sum, frequently carry out statistics find e to word, the frequency that d occurs is higher, temperature can be thought on the whole in downward trend, by finding that to the statistics of time these changes have more now: during October 18 to November 14, also meet the change of the natural law.
In the temperature data of the level of tail water, directed line segment representated by the limit that line thicker in complex network is corresponding is: ee → ee, de → ee, ee → ed, ee → er, er → de, these changes account for 71.8% of change sum, frequently carry out statistics find e to word, the frequency that d occurs is higher, temperature can be thought on the whole in downward trend, by finding that to the statistics of time these changes have more now: during October 18 to November 14, also meet the change of the natural law.
And the higher node of the thermometer data interior joint degree of dam body has dd (308), ff (213), tt (207), ee (185), can see that egress randomness is stronger, there is no relatively uniform variation tendency, carry out statistics to the version of node to find, ff → ff, several change frequencies such as dd → dd, dd → de are maximum, but these changes only account for 7.8% of total change number, thus can think that these group data are all not representative, the randomness of this network is stronger.
Analyze respectively their topological property below, as shown in table 14 is upstream and downstream water temperature and temperature complex network index.
Table 14
Analysis is compared by group data each in table 14, obviously, upstream temperature data have more representativeness, and the temperature data of upstream has less modularity coefficient simultaneously, prove that the correlation comparison between its inner each group variety is sparse, therefore can think that these group data do not possess obvious community structure.Comprehensive above analysis, the packet of upstream temperature, containing more information, should carry out the monitoring of emphasis in daily monitoring.
Innovative point of the present invention is, the Complex Networks Analysis method of time series feature adopts coarse method, is conducive to overcoming the impact of data error on analysis result, is conducive to the analysis realizing different scale; Adopt Complex Networks Analysis method, the significance level of each measuring point can be found, be conducive to analyzing the correlationship between each measuring point, and responsive measuring point can be extracted, thus the healthy condition of the engineering grasping dykes and dams timely and accurately, be conducive to realizing the more effective early warning of dykes and dams condition.
More than show and describe ultimate principle of the present invention, principal character and advantage.The technician of the industry should understand; the present invention is not restricted to the described embodiments; what describe in above-described embodiment and instructions just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.Application claims protection domain is defined by appending claims and equivalent thereof.
Claims (7)
1., based on the responsive measuring point analytical approach of dykes and dams health monitoring of complex network, it is characterized in that, comprise the following steps:
1) according to dykes and dams health monitoring project and measuring point, application is laid in the actual measured value time series of the corresponding measuring point of sensor acquisition dykes and dams of measuring point;
2) use the method for equiprobability coarse, measured value time series is changed into symbol sebolic addressing;
3) symbol sebolic addressing is coded in the topological structure of complex network, complex structure networking;
4) analyze the characteristic parameter of complex network, obtain responsive measuring point and the condition Evolution thereof of this dykes and dams health monitoring project, the healthy condition of engineering of carrying out dykes and dams differentiates.
2. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 1, is characterized in that: described step 1) in dykes and dams health monitoring project comprise environment parameter, seepage field, temperature field and deformation field.
3. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 1, is characterized in that: described step 2) method of equal probability coarse, specifically comprise step,
2-1) establish x
maxand x
minbe measured value seasonal effect in time series maximal value and minimum value respectively, the symbol after conversion has N kind, is designated as s
1, s
2..., s
n, definition
S
i＝s
j，x
min+(j-1)d≤x
i＜x
min+jd，j＝1，…，N(1)
Wherein, N>2, d=(x
max-x
min)/N; The symbol sebolic addressing that N kind kinds of characters forms is obtained by formula (1);
2-2) remember dykes and dams measured value time series for x (t)=1 ..., N}, through type (2) calculates measured value time series fluctuation k (t),
Wherein, Δ t is the time interval;
2-3) through type (3) calculates the Probability p (k) that different undulating quantity may occur,
Wherein, the number of times of the fluctuation mode x generation of the corresponding measured value sequence of Num (x);
2-4) measured value seasonal effect in time series fluctuation k (t) is divided into 5 intervals, defining 5 characteristic characters is formula (4),
Wherein, t represents measured value numerical value to be increased fast, and r represents measured value numerical value to be increased slowly, and e represents measured value numerical value and do not increase and do not subtract, and d represents measured value numerical value to be reduced slowly, and f represents measured value numerical value to be reduced fast;
Thus measured value time series to be converted to symbol sebolic addressing be formula (5),
S＝(S
1S
2S
3…)，S
i∈(t，r，e，d，f)(5)。
4. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 1, it is characterized in that: described step 3) middle complex structure networking, be after the topological structure of complex network symbol sebolic addressing being coded in oriented weighting, convert the accessible form of pajek software to by excle2pajek software.
5. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 1, is characterized in that: described step 4) in the characteristic parameter of complex network is analyzed, undertaken by pajek software.
6. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 1, is characterized in that: described step 4) in the characteristic parameter of complex network comprise average path length, on average gather coefficient and modularity coefficient.
7. the responsive measuring point analytical approach of the dykes and dams health monitoring based on complex network according to claim 2, is characterized in that: described environment parameter comprises level of tail water measuring point and atmospheric pressure measuring point before upper pond level measuring point before dam break, dam break;
Described seepage field comprises osmotic pressure water level measuring point, seepage flow measuring point and dam body moisture measuring point;
Described temperature field adopts distribution type fiber-optic to measure, and comprises dam body surface temperature measuring point, storehouse water temperature measuring point, dam body internal temperature measuring point, upstream water temperature measuring point and downstream water temperature measuring point;
Described deformation field comprises dam body surface deformation measuring point, dam body internal modification measuring point and inclination measuring point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510691360.5A CN105404709B (en) | 2015-10-22 | 2015-10-22 | Dykes and dams health monitoring sensitivity measuring point analysis method based on complex network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510691360.5A CN105404709B (en) | 2015-10-22 | 2015-10-22 | Dykes and dams health monitoring sensitivity measuring point analysis method based on complex network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105404709A true CN105404709A (en) | 2016-03-16 |
| CN105404709B CN105404709B (en) | 2018-10-02 |
Family
ID=55470196
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510691360.5A Active CN105404709B (en) | 2015-10-22 | 2015-10-22 | Dykes and dams health monitoring sensitivity measuring point analysis method based on complex network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105404709B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944613A (en) * | 2017-11-17 | 2018-04-20 | 湖南大学 | A kind of hazard forecasting method based on symbolic network |
| CN110069832A (en) * | 2019-04-01 | 2019-07-30 | 南京航空航天大学 | Discontinuous plate malformation inverting and joining method based on geometric coordinate transfer algorithm |
| CN111047079A (en) * | 2019-11-25 | 2020-04-21 | 山东师范大学 | Wind power plant wind speed time series prediction method and system |
| CN113837674A (en) * | 2021-11-29 | 2021-12-24 | 江西省水利科学院 | Dyke dangerous case sensitivity analysis method based on weight model |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915555A (en) * | 2015-05-28 | 2015-09-16 | 水利部南京水利水文自动化研究所 | Dam break grading early-warning index extraction method for homogeneous earth dam |
-
2015
- 2015-10-22 CN CN201510691360.5A patent/CN105404709B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915555A (en) * | 2015-05-28 | 2015-09-16 | 水利部南京水利水文自动化研究所 | Dam break grading early-warning index extraction method for homogeneous earth dam |
Non-Patent Citations (6)
| Title |
|---|
| SUIGEWU: "《如何将excel形式数据转化为pajek的数据形式》", 《HTTPS://WWW.DOUBAN.COM/NOTE/200076634/》 * |
| 刘典勇等: "《基于Pajek的航空网络特征评价与可生存性分析》", 《电子科技》 * |
| 周磊等: "《利用复杂网络研究中国温度序列的拓扑性质》", 《物理学报》 * |
| 唐强等: "《径流时间序列复杂网络的动力学拓扑性质》", 《武汉纺织大学学报》 * |
| 张佃中等: "《基于等概率粗粒化的复杂度算法及其应用》", 《系统仿真学报》 * |
| 林聚任: "《社会网络分析：理论、方法与应用"》", 30 April 2009 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944613A (en) * | 2017-11-17 | 2018-04-20 | 湖南大学 | A kind of hazard forecasting method based on symbolic network |
| CN110069832A (en) * | 2019-04-01 | 2019-07-30 | 南京航空航天大学 | Discontinuous plate malformation inverting and joining method based on geometric coordinate transfer algorithm |
| CN111047079A (en) * | 2019-11-25 | 2020-04-21 | 山东师范大学 | Wind power plant wind speed time series prediction method and system |
| CN111047079B (en) * | 2019-11-25 | 2022-05-13 | 山东师范大学 | Wind power plant wind speed time series prediction method and system |
| CN113837674A (en) * | 2021-11-29 | 2021-12-24 | 江西省水利科学院 | Dyke dangerous case sensitivity analysis method based on weight model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105404709B (en) | 2018-10-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| D'Andrade | U-statistic hierarchical clustering | |
| CN105404709A (en) | Complex network based sensitive monitoring point analysis method for dam health monitoring | |
| CN109193650B (en) | Power grid weak point evaluation method based on high-dimensional random matrix theory | |
| CN111277332A (en) | Performance state detection method and device of optical module in engineering application and electronic equipment | |
| CN110213003B (en) | Wireless channel large-scale fading modeling method and device | |
| CN104915555B (en) | The extracting method of homogeneous earth dam dam break grading forewarning system index | |
| Sahu et al. | Regional frequency analysis using L-moment methodology—a review | |
| CN105843733B (en) | A kind of method for testing performance and device of big data platform | |
| CN107483487B (en) | TOPSIS-based multi-dimensional network security measurement method | |
| Ahmadi et al. | Input selection for long-lead precipitation prediction using large-scale climate variables: a case study | |
| CN104049001B (en) | Vertical profit phase content based on frequency complex network is measured and verification method | |
| CN110389971A (en) | A kind of multi-Sensor Information Fusion Approach based on cloud computing | |
| CN110321493A (en) | A kind of abnormality detection of social networks and optimization method, system and computer equipment | |
| Momtazpour et al. | Analyzing invariants in cyber-physical systems using latent factor regression | |
| CN104111887A (en) | Software fault prediction system and method based on Logistic model | |
| Han et al. | Generating uncertain networks based on historical network snapshots | |
| Deng et al. | Spatial–temporal pattern and influencing factors of drought impacts on agriculture in China | |
| CN105722129A (en) | Wireless sensing network event detection method and system based on FSAX-MARKOV model | |
| CN109460571A (en) | A kind of smart grid data reconstruction method based on matrix low-rank and sparsity | |
| CN112507591A (en) | Gravity dam crest displacement monitoring method based on separation type hybrid modeling technology | |
| CN109145257A (en) | A kind of dam safety monitoring analysis method based on the public abnormal time of multi-measuring point | |
| Bai et al. | Evolving an information diffusion model using a genetic algorithm for monthly river discharge time series interpolation and forecasting | |
| Li et al. | Improved river water-stage forecasts by ensemble learning | |
| CN116027158A (en) | High-voltage cable partial discharge fault prediction method and system | |
| CN106550387B (en) | A kind of wireless sensor network routing layer QoS evaluating method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |