CN107451622A - A kind of tunnel operation state division methods based on big data cluster analysis - Google Patents
A kind of tunnel operation state division methods based on big data cluster analysis Download PDFInfo
- Publication number
- CN107451622A CN107451622A CN201710710261.6A CN201710710261A CN107451622A CN 107451622 A CN107451622 A CN 107451622A CN 201710710261 A CN201710710261 A CN 201710710261A CN 107451622 A CN107451622 A CN 107451622A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mtd
- munderover
- cluster
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Abstract
The present invention relates to Tunnel Engineering field, and in particular to a kind of tunnel operation state division methods based on big data cluster analysis.Its ground adapts to tunnel operation state dynamic change and requires and provide accurate prediction.The method and step that the present invention uses for:1):Tunnel operation monitoring experiment is carried out, obtains vcehicular tunnel operation state hyperspace;2):Monitor sample data set is pre-processed, then judged whether containing missing values and noise data, all abnormal datas are rejected;3):Its any sample x is calculated to the sample set for being divided into k clusteriSilhouette coefficient, and then calculate it is different k cluster under all sample n mean profile coefficient SkAnd assert and work as SkIt is tunnel operation state optimum state classification number to take k values corresponding during maximum;4):Preset cluster number is k, and application FCM algorithms run monitor sample data set to tunnel and carry out cluster analysis, obtains degree of membership and k cluster centre of all samples to k cluster;5):According to degree of membership maximum principle, it is determined that operation state classification corresponding to each tunnel operation monitor sample.
Description
Technical field
The present invention relates to Tunnel Engineering field, and in particular to a kind of tunnel operation state based on big data cluster analysis is drawn
Divide method.
Background technology
A large amount of Extra-long Highway Tunnels are built up and put into effect successively in recent years, and vcehicular tunnel is gradually turned by building peak period
To operation peak period.But due to being influenceed by traffic composition and the volume of traffic, pollutant continues to build up in hole, discharge or dilute relatively tired
It is difficult so that ventilating problem turns into the matter of utmost importance faced during operation, and problem is brought to tunnel operation and management.Therefore, it is necessary to right
Traffic flow data is analyzed with environmental monitoring data in tunnel, it is determined that after tunnel operation state, formulates corresponding operation pipe
Control measure.The wherein reasonability of tunnel operation state division methods and science, directly determine tunnel operation management and control measures
Validity.But because Extra-long Highway Tunnel operation state is by the traffic such as the people in tunnel, car, road, environment key element mutual shadow jointly
Ring superposition to produce, its influence factor is numerous, Evolution is complicated, there is no the division methods of science and unified division mark at present
It is accurate.Also real-time traffic stream information and ventilated environment information in comprehensive utilization tunnel is had no in published document and patent document to enter
The research invention of row operation state analysis.
Cluster analysis is exactly according to sample distribution in space, incorporates into closely located sample into one kind, its principle is to make
Between class distance it is as big as possible, distance is as small as possible between sample in class, so as to obtain the division methods of a variety of operation states.It is fuzzy
C- averages (Fuzzy C-Means, FCM) algorithm is a kind of extensive subarea clustering method, earliest by Dunn propose and by
Bezdek is promoted, and is taken advantage in big data application aspect.On the basis of FCM algorithms, Kaufman et al. proposes one kind
New fuzzy clustering algorithm-FANNY algorithms.Compared with general FCM algorithms, FANNY algorithms are sensitive to wrong data or exceptional value
Spend relatively low, while have more preferable identification capability to aspherical cluster.
Therefore, the method that cluster analysis can be used, arithmetic for real-time traffic flow and the big number of environmental monitoring during being runed for tunnel
According to carrying out integrating computing, a kind of tunnel operation state division methods based on big data cluster analysis are thus proposed, so as to more preferable
Ground adapts to tunnel operation state dynamic change and requires and provide accurate prediction, to formulate Extra-long Highway Tunnel ventilation equipment and traffic
Run intelligent management and control scheme and theoretical foundation and scientific method are provided.
The content of the invention
In view of this, the present invention provides a kind of tunnel operation state division methods based on big data cluster analysis.
To solve the problems, such as that prior art is present, the technical scheme is that:It is a kind of based on big data cluster analysis
Tunnel operation state division methods, it is characterised in that:Described method and step is:
Step 1):Tunnel operation monitoring is carried out first:
Gather all kinds of pollutant concentrations, visibility, wind speed, traffic flow data structure tunnel operation monitor sample data
Collection, obtains vcehicular tunnel operation state hyperspace;
Step 2):Monitor sample data set is pre-processed:
Hough transformation is carried out first:Or weak related dimension uncorrelated to tunnel operation state is deleted, is then judged whether
Containing missing values and noise data, all abnormal datas are rejected;
Step 3):Cluster analysis is carried out according to the monitor sample data set after processing, makes its between class distance for maximum, and class
Distance is minimum between interior sample, using fuzzy clustering algorithm-FANNY algorithms, is divided into k and clusters and calculate its any sample xi
Silhouette coefficient Sk(xi), and then calculate the mean profile coefficient S of all sample n under different k clusterskAnd assert and work as SkTake most
Corresponding k values are tunnel operation state optimum state classification number during big value;
Step 4):Preset cluster number is k, and application FCM algorithms cluster to tunnel operation monitor sample data set
Analysis, obtain degree of membership and k cluster centre of all samples to k cluster;
Step 5):According to degree of membership maximum principle, it is determined that operation state classification corresponding to each tunnel operation monitor sample.
2nd, a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its
It is characterised by:Any sample x of its in the step 3)iSilhouette coefficient Sk(xi) calculation formula be:
In formula (1):A(xi) represent sample xiWith the average dissimilarity of other samples in its affiliated cluster;If Dc(xi) represent
Sample xiAverage dissimilarity with clustering c samples in other k-1 clustering cluster, it is B (x to take minimum value in all k-1 valuesi):
In formula (2):It is B (x to take minimum value in all k-1 valuesi)。
3rd, a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its
It is characterised by:All sample n mean profile coefficient S under the different k cluster of the step 3)kComputational methods be:
In formula (3):The mean profile coefficient under different k cluster numbers on all sample n is calculated with mean profile method
Sk。
4th, a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its
It is characterised by:
The step 4) assumes i-th of observation xi=(xi1,xi2,xi3,xi4,xi5), CO, NO are represented respectively2, it is wind speed, thin
Five particulate matter, HGV dimension monitoring results, the sample data set of n observation are denoted as X, and X can use the matrix of N × 5 to represent, i.e.,:
FCM algorithms are based on minimizing following object function:
In formula (5):K represents preset operation state number, that is, clusters number;υ represents cluster numbering;mυRepresent cluster υ's
Center;Represent sample xiTo cluster centre mυSubjection degree, its Fuzzy Exponential be 2;||xi-mυ||2Represent sample xiWith mυ
Between Euclidean distance square;F represents the dimension numbering of operation state quintuple space, mυ1、mυ2、mυ3、mυ4And mυ5Represent respectively poly-
Class center mυCorresponding CO, NO2, wind speed, fine particle, HGV values.mυfCalculated by following formula:
On the basis of FCM algorithms, a kind of new fuzzy clustering FANNY algorithms can be derived by formula (5), such as formula (7) table
Show:
Its constraints is:
uiυ>=0, i=1 ..., N;
Successive ignition is carried out to data after pretreatment according to FANNY algorithms in formula (6) and formula (7), obtains three kinds of tunnel fortune
The cluster centre and sample of battalion's state are subordinate to angle value u for each clusteriυ(1≤i≤N,1≤υ≤k)。
Compared with prior art, advantages of the present invention is as follows:
1) the inventive method is dense to the traffic flow modes such as traffic composition, magnitude of traffic flow data in tunnel and each pollutant
The environmental status datas such as degree, wind speed are merged, and run Monitoring Data collection by analyzing tunnel, and then run management and control for tunnel and carry
For real-time scientific basis;
2) advantage that the present invention is handled using cluster algorithm for big data, tunnel operation is perceived by data-driven
State, so as to better adapt to tunnel operation change requirement and provide accurate prediction, and then improve tunnel operation management and control precision;
3) present invention applies FCM clustering algorithms, to take average silhouette coefficient maximum as principle, so that it is determined that tunnel operation
The optimal cluster number k of state, and then ensure Clustering Effect accuracy and science.
Brief description of the drawings:
Fig. 1 is the step flow chart of the inventive method;
Fig. 2 is Qinling Mountains No.1 tunnel south provided in an embodiment of the present invention line structural representation;
Fig. 3 is that Monitoring Data explanation figure is runed in tunnel provided in an embodiment of the present invention;
Fig. 4 is that mean profile coefficient provided in an embodiment of the present invention calculates figure.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Tunnel operation state classification problem self-defined first:Given training set T={ (x1,y1),…,(xN,yN)}∈(X5×
Y)N, wherein xi∈X5I-th of sample in the tunnel operation monitor sample data set of presentation class mode input, includes CO, NO2、
The monitoring results such as wind speed, fine particle, traction engine (Heavy Goods Vehicles, HGV), yi∈ Y={ c1,c2,c3,c4Table
One kind of four kinds of states such as slight pollution, intermediate pollution, serious pollution and serious pollution, i=1 ..., N tables corresponding to sample sheet
Show sample number in training set, find state space X accordingly5On a decision function f (x):X5→ Y, to infer any prison
Tunnel operation state corresponding to test sample sheet.From definition, tunnel operation state classification problem is solved, is substantially exactly to find handle
N dimension state spaces XnIt is divided into the rule of some.
A kind of tunnel operation state division methods step based on big data cluster analysis is:
Step 1):Tunnel operation monitoring is carried out first:
Gather the data such as all kinds of pollutant concentrations, visibility, wind speed, magnitude of traffic flow structure tunnel operation monitor sample data
Collection, obtains vcehicular tunnel operation state hyperspace;
Step 2):Monitor sample data set is pre-processed:
Hough transformation is carried out first:Or weak related dimension uncorrelated to tunnel operation state is deleted, is then judged whether
Containing missing values and noise data, all abnormal datas are rejected;
The main reason for causing missing values has gas monitor apparatus or wagon detector failure, can not normally detect.Specifically
The missing in space is shown as, clearly can be found out from data.Noise data mainly includes pollutant concentration or the volume of traffic surpasses
Zone of reasonableness etc. is crossed, it is suddenly big or suddenly small.
Step 3):Cluster analysis is carried out according to the monitor sample data set after processing, makes its between class distance as big as possible, and
Distance is as small as possible between sample in class, using fuzzy clustering algorithm-FANNY algorithms, is divided into k and clusters and to calculate its any
Sample xiSilhouette coefficient Sk(xi), and then calculate the mean profile coefficient S of all sample n under different k clusterskAnd assert and work as Sk
It is tunnel operation state optimum state classification number to take k values corresponding during maximum;
Its any sample xiSilhouette coefficient Sk(xi) calculation formula be:
In formula (1):A(xi) represent sample xiWith the average dissimilarity of other samples in its affiliated cluster;If Dc(xi) represent
Sample xiAverage dissimilarity with clustering c samples in other k-1 clustering cluster, it is B (x to take minimum value in all k-1 valuesi):
B(xi)=minD (xi) (9)
In formula (2):It is B (x to take minimum value in all k-1 valuesi)。
All sample n mean profile coefficient S under different k clusterskComputational methods be:
In formula (3):The mean profile coefficient under different k cluster numbers on all sample n is calculated with mean profile method
Sk。
Step 4):Preset cluster number is k, and application FCM algorithms cluster to tunnel operation monitor sample data set
Analysis, obtain degree of membership and k cluster centre of all samples to k cluster;
The step 4) assumes i-th of observation xi=(xi1,xi2,xi3,xi4,xi5), CO, NO are represented respectively2, it is wind speed, thin
Five dimension monitoring results such as particulate matter, HGV, the sample data set of n observation are denoted as X, and X can use the matrix of N × 5 to represent,
I.e.:
FCM algorithms are based on minimizing following object function:
In formula (5):K represents preset operation state number, that is, clusters number;υ represents cluster numbering;mυRepresent cluster υ's
Center;Subjection degrees of the sample xi to cluster centre m υ is represented, its Fuzzy Exponential is 2;||xi-mυ||2Represent sample xiWith mυ
Between Euclidean distance square;F represents the dimension numbering of operation state quintuple space, mυ1、mυ2、mυ3、mυ4And mυ5Represent respectively poly-
Class center mυCorresponding CO, NO2, wind speed, fine particle, HGV values.mυfCalculated by following formula:
On the basis of FCM algorithms, a kind of new fuzzy clustering FANNY algorithms can be derived by formula (5), such as formula (7) table
Show:
Its constraints is:
uiυ>=0, i=1 ..., N;
Successive ignition is carried out to data after pretreatment according to FANNY algorithms in formula (6) and formula (7), obtains three kinds of tunnel fortune
The cluster centre and sample of battalion's state are subordinate to angle value u for each clusteriυ(1≤i≤N,1≤υ≤k)。
Step 5):According to degree of membership maximum principle, it is determined that operation state classification corresponding to each tunnel operation monitor sample.
Embodiment
Chinese highway Qinling Mountains No.1 super long tunnel relies on for engineering to the west of the present invention, carries out tunnel operation monitoring experiment simultaneously
The embodiment of the present invention is described further with reference to accompanying drawing.
As shown in figure 1, a kind of tunnel operation state division methods based on cluster analysis, are specifically included:Tunnel operation prison
Survey the steps such as data prediction, optimal cluster number differentiation, cluster analysis, tunnel operation state division result.
As shown in Fig. 2 Qinling Mountains No.1 tunnel is the double hole highway tunnel with four lanes of separate type, and southern line total length 6102m, tunnel portal
Height above sea level is 1322m, and outlet height above sea level is 1391m, and average gradient is+2.58%, sets (ESA-1 to ESA- at the band 11 that stops in emergency
11), jet blower 30 is installed altogether, reserved at plenum ventilation inclined shaft 1.Because the tunnel uses full ventilation by force draft side
Formula, the pollutant concentration in tunnel meet " upper triangle " distribution characteristics, that is, it is minimum to enter hole concentration, goes out hole concentration highest.It is comprehensive
Upper analysis, it is final to choose the monitoring place that lay-by is studied as the present invention at the ESA-11 for polluting most serious.
As shown in figure 3, collection tunnel environment data and the item data of two major class of traffic census seven, wherein traffic census will
Car model classification is passenger car (Passenger Cars, PC), Light-duty Vehicle (Light-Duty Vehicles, LDV) and traction engine
Three kinds of (Heavy-Goods Vehicles, HGV) etc..In traffic composition within the tunnel operation monitoring phase, PC, LDV and HGV ratio
Example is respectively that 29.46%, 3.21% and 67.32%, LDV accounting are relatively low, therefore negligible influences of the LDV to operation state.Pass through
Pearson correlation coefficient is understood between calculating variable, CO, NO2It is in strong correlation relation with HGV, is in weak dependency relation with PC, therefore
Also influences of the PC to operation state can be neglected, only retain HGV.Therefore finally choose CO, NO2, wind speed, fine particle, five classes such as HGV
The sample data set that data are studied as the present invention.
Rejecting outliers are carried out to the sample data set of monitoring, judged whether containing missing values and noise data.Cause to lack
The main reason for mistake value is gas monitor apparatus or wagon detector failure, and noise data is mainly pollutant concentration or the volume of traffic
Beyond zone of reasonableness.Due to exceptional value, to account for total sample proportion smaller, therefore all abnormal datas are rejected.
As shown in figure 4, before carrying out cluster analysis to sample data set, first determined according to mean profile coefficient principle optimal
Number is clustered, the sample set to being divided into k cluster, its any sample xiSilhouette coefficient calculates according to following formula:
In formula (1):A(xi) represent sample xiWith the average dissimilarity of other samples in its affiliated cluster;If Dc(xi) represent
Sample xiAverage dissimilarity with clustering c samples in other k-1 clustering cluster.It is B (x to take minimum value in all k-1 valuesi):
B(xi)=minD (xi) (16)
In formula (2):It is B (x to take minimum value in all k-1 valuesi)。
In formula (3):The mean profile coefficient under different k cluster numbers on all sample n is calculated with mean profile method
Sk.As shown in Figure 4 as k=3, mean profile coefficient SkMaximum, now Clustering Effect is best.Therefore for being adopted in the present embodiment
It is 3 that the tunnel operation monitor sample data set of collection, which takes optimal operation state number,.
In the present embodiment, monitor sample data set is runed according to tunnel and obtains vcehicular tunnel operation state quintuple space,
By the sample distribution in space, closely located sample is incorporated into one kind.Using FCM clustering methods, it is assumed that i-th of observation
It is worth for xi=(xi1,xi2,xi3,xi4,xi5), CO, NO are represented respectively2, wind speed, fine particle, five dimension monitoring results such as HGV,
The sample data set of n observation is denoted as X, and X can use the matrix of N × 5 to represent, i.e.,:
FCM algorithms are based on minimizing following object function:
In formula (5):K represents preset operation state number, that is, clusters number;υ represents cluster numbering;mυRepresent cluster υ's
Center;Represent sample xiTo cluster centre mυSubjection degree, its Fuzzy Exponential be 2;||xi-mυ||2Represent sample xiWith mυ
Between Euclidean distance square;F represents the dimension numbering of operation state quintuple space, mυ1、mυ2、mυ3、mυ4And mυ5Represent respectively poly-
Class center mυCorresponding CO, NO2, wind speed, fine particle, HGV values.mυfCalculated by following formula:
On the basis of FCM algorithms, a kind of new fuzzy clustering FANNY algorithms can be derived by formula (5), such as formula (7) table
Show:
Its constraints is:
uiυ>=0, i=1 ..., N;
In the present embodiment, successive ignition is carried out to data after pretreatment according to FANNY algorithms in formula (6) and formula (7), obtained
Cluster centre and sample to three kinds of tunnel operation states are subordinate to angle value u for each clusteriυ(1≤i≤N,1≤υ≤k)。
Correspond to CO (ppm), the NO of each cluster centre in formula (8) matrix respectively per a line2(ppm), wind speed (m/s), thin
Grain thing (mg/m3) and HGV (veh/15min) value.It is 0 according to the first row HGV, the second row, the third line NO2More than 1ppm deducibilitys
Go out it and correspond to slight (c respectively1), severe (c3) and serious (c4) three kinds of tunnel operation state cluster centres are polluted, the volume of traffic is smaller
When moderate (c2) pollutional condition missing.
According to degree of membership maximum principle, sample generic is determined, so as to complete the division of tunnel operation state.
As shown in table 1, according to tunnel operation state division result, corresponding ventilation equipment and traffic circulation management and control can be formulated
Strategy, and then a whole set of management and control scheme is formed, so as to lift Extra-long Highway Tunnel operation regulatory level.
Table 1
Tunnel operation state | Ventilation equipment management and control strategy | Traffic circulation management and control strategy |
Slight pollution | Gravity-flow ventilation | Nothing |
Serious pollution | Jet blower | Nothing |
Serious pollution | Jet blower+axial flow blower | Consider that limitation HGV passes through |
Above content is the further description made in conjunction with specific embodiments to the inventive method, it is impossible to assert this hair
The specific implementation of bright method is only limited to this.For general technical staff of the technical field of the invention, this is not being departed from
Some equivalent substitutes or obvious modification are made on the premise of inventive concept, and performance or purposes are identical, should all be considered as belonging to this
The scope of patent protection that invention is determined by the claims submitted.
Claims (4)
- A kind of 1. tunnel operation state division methods based on big data cluster analysis, it is characterised in that:Described method and step For:Step 1):Tunnel operation monitoring is carried out first:All kinds of pollutant concentrations, visibility, wind speed, traffic flow data structure tunnel operation monitor sample data set are gathered, is obtained To vcehicular tunnel operation state hyperspace;Step 2):Monitor sample data set is pre-processed:Hough transformation is carried out first:Delete or weak related dimension uncorrelated to tunnel operation state, then judge whether containing Missing values and noise data, all abnormal datas are rejected;Step 3):Cluster analysis is carried out according to the monitor sample data set after processing, makes its between class distance for maximum, and sample in class Distance is minimum between this, using fuzzy clustering algorithm-FANNY algorithms, is divided into k and clusters and calculate its any sample xiWheel Wide coefficient Sk(xi), and then calculate the mean profile coefficient S of all sample n under different k clusterskAnd assert and work as SkTake maximum When corresponding k values classify number for tunnel operation state optimum state;Step 4):Preset cluster number is k, and application FCM algorithms run monitor sample data set to tunnel and carry out cluster analysis, Obtain degree of membership and k cluster centre of all samples to k cluster;Step 5):According to degree of membership maximum principle, it is determined that operation state classification corresponding to each tunnel operation monitor sample.
- 2. a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its feature It is:Any sample x of its in the step 3)iSilhouette coefficient Sk(xi) calculation formula be:<mrow> <msub> <mi>S</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mi>B</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <msub> <mi>A</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> <mrow> <mo>(</mo> <msub> <mi>A</mi> <mi>k</mi> </msub> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> <mo>,</mo> <msub> <mi>B</mi> <mi>k</mi> </msub> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>In formula (1):A(xi) represent sample xiWith the average dissimilarity of other samples in its affiliated cluster;If Dc(xi) represent sample xiAverage dissimilarity with clustering c samples in other k-1 clustering cluster, it is B (x to take minimum value in all k-1 valuesi):B(xi)=minD (xi) (2)In formula (2):It is B (x to take minimum value in all k-1 valuesi)。
- 3. a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its feature It is:All sample n mean profile coefficient S under the different k cluster of the step 3)kComputational methods be:<mrow> <msub> <mi>S</mi> <mi>k</mi> </msub> <mo>=</mo> <mfrac> <mn>1</mn> <mi>N</mi> </mfrac> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>S</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>)</mo> </mrow> <mo>,</mo> <mi>k</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>...</mo> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>In formula (3):The mean profile coefficient S under different k cluster numbers on all sample n is calculated with mean profile methodk。
- 4. a kind of tunnel operation state division methods based on big data cluster analysis according to claim 1, its feature It is:The step 4) assumes i-th of observation xi=(xi1,xi2,xi3,xi4,xi5), CO, NO are represented respectively2, wind speed, fine grained Five thing, HGV dimension monitoring results, the sample data set of n observation are denoted as X, and X can use the matrix of N × 5 to represent, i.e.,:<mrow> <mi>X</mi> <mo>=</mo> <mfenced open = "[" close = "]"> <mtable> <mtr> <mtd> <msub> <mi>x</mi> <mn>11</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>12</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>13</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>14</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>15</mn> </msub> </mtd> </mtr> <mtr> <mtd> <msub> <mi>x</mi> <mn>21</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>22</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>23</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>24</mn> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mn>25</mn> </msub> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> <mtd> <mo>.</mo> </mtd> </mtr> <mtr> <mtd> <msub> <mi>x</mi> <mrow> <mi>N</mi> <mn>1</mn> </mrow> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mrow> <mi>N</mi> <mn>2</mn> </mrow> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mrow> <mi>N</mi> <mn>3</mn> </mrow> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mrow> <mi>N</mi> <mn>4</mn> </mrow> </msub> </mtd> <mtd> <msub> <mi>x</mi> <mrow> <mi>N</mi> <mn>5</mn> </mrow> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow>FCM algorithms are based on minimizing following object function:<mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>&upsi;</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> <mo>|</mo> <mo>|</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>m</mi> <mi>&upsi;</mi> </msub> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> <mo>=</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>&upsi;</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> <munderover> <mo>&Sigma;</mo> <mrow> <mi>f</mi> <mo>=</mo> <mn>1</mn> </mrow> <mn>5</mn> </munderover> <msup> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mrow> <mi>i</mi> <mi>f</mi> </mrow> </msub> <mo>-</mo> <msub> <mi>m</mi> <mrow> <mi>&upsi;</mi> <mi>f</mi> </mrow> </msub> <mo>)</mo> </mrow> <mn>2</mn> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow>In formula (5):K represents preset operation state number, that is, clusters number;υ represents cluster numbering;mυRepresent cluster υ center;Represent sample xiTo cluster centre mυSubjection degree, its Fuzzy Exponential be 2;||xi-mυ||2Represent sample xiWith mυBetween Euclidean Square of distance;F represents the dimension numbering of operation state quintuple space, mυ1、mυ2、mυ3、mυ4And mυ5Cluster centre m is represented respectivelyυ Corresponding CO, NO2, wind speed, fine particle, HGV values.mυfCalculated by following formula:<mrow> <msub> <mi>m</mi> <mrow> <mi>&upsi;</mi> <mi>f</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> <msub> <mi>x</mi> <mrow> <mi>i</mi> <mi>f</mi> </mrow> </msub> </mrow> <mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>On the basis of FCM algorithms, a kind of new fuzzy clustering FANNY algorithms can be derived by formula (5), as formula (7) represents:<mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> <munderover> <mo>&Sigma;</mo> <mrow> <mi>&upsi;</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <mfrac> <mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <munderover> <mo>&Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> <msubsup> <mi>u</mi> <mrow> <mi>j</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> <mo>|</mo> <mo>|</mo> <msub> <mi>x</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>x</mi> <mi>j</mi> </msub> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> </mrow> <mrow> <mn>2</mn> <munderover> <mo>&Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>u</mi> <mrow> <mi>j</mi> <mi>&upsi;</mi> </mrow> <mn>2</mn> </msubsup> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>Its constraints is:uiυ>=0, i=1 ..., N;<mrow> <munderover> <mo>&Sigma;</mo> <mrow> <mi>&upsi;</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <msub> <mi>u</mi> <mrow> <mi>i</mi> <mi>&upsi;</mi> </mrow> </msub> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mo>...</mo> <mo>,</mo> <mi>N</mi> <mo>;</mo> </mrow>Successive ignition is carried out to data after pretreatment according to FANNY algorithms in formula (6) and formula (7), obtains three kinds of tunnel operation shapes The cluster centre and sample of state are subordinate to angle value u for each clusteriυ(1≤i≤N,1≤υ≤k)。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710710261.6A CN107451622A (en) | 2017-08-18 | 2017-08-18 | A kind of tunnel operation state division methods based on big data cluster analysis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710710261.6A CN107451622A (en) | 2017-08-18 | 2017-08-18 | A kind of tunnel operation state division methods based on big data cluster analysis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107451622A true CN107451622A (en) | 2017-12-08 |
Family
ID=60492612
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710710261.6A Pending CN107451622A (en) | 2017-08-18 | 2017-08-18 | A kind of tunnel operation state division methods based on big data cluster analysis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107451622A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108564110A (en) * | 2018-03-26 | 2018-09-21 | 上海电力学院 | A kind of Air Quality Forecast method based on clustering algorithm |
CN109213752A (en) * | 2018-08-06 | 2019-01-15 | 国网福建省电力有限公司信息通信分公司 | A kind of data cleansing conversion method based on CIM |
CN109374631A (en) * | 2018-09-30 | 2019-02-22 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | A kind of tunnel state evaluating method |
CN109993555A (en) * | 2017-12-30 | 2019-07-09 | 中国移动通信集团四川有限公司 | Internet television potential user complains prediction technique, device and equipment |
CN110705602A (en) * | 2019-09-06 | 2020-01-17 | 平安科技(深圳)有限公司 | Large-scale data clustering method and device and computer readable storage medium |
CN113447813A (en) * | 2020-09-03 | 2021-09-28 | 鲁能集团有限公司 | Fault diagnosis method and equipment for offshore wind generating set |
CN114821514A (en) * | 2022-06-30 | 2022-07-29 | 交通运输部公路科学研究所 | A remove detecting system for detecting tunnel road surface wet and slippery state |
CN116611620A (en) * | 2023-07-18 | 2023-08-18 | 厚德智能技术(山东)有限公司 | Smart city safety collaborative management information system |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150032759A1 (en) * | 2012-04-06 | 2015-01-29 | Sk Planet Co., Ltd. | System and method for analyzing result of clustering massive data |
CN105809272A (en) * | 2016-02-25 | 2016-07-27 | 大连理工大学 | Step hydropower station group instruction scheduling optimization method based on data mining |
CN106297326A (en) * | 2016-10-27 | 2017-01-04 | 深圳榕亨实业集团有限公司 | Based on holographic road network tide flow stream Lane use control method |
-
2017
- 2017-08-18 CN CN201710710261.6A patent/CN107451622A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150032759A1 (en) * | 2012-04-06 | 2015-01-29 | Sk Planet Co., Ltd. | System and method for analyzing result of clustering massive data |
CN105809272A (en) * | 2016-02-25 | 2016-07-27 | 大连理工大学 | Step hydropower station group instruction scheduling optimization method based on data mining |
CN106297326A (en) * | 2016-10-27 | 2017-01-04 | 深圳榕亨实业集团有限公司 | Based on holographic road network tide flow stream Lane use control method |
Non-Patent Citations (1)
Title |
---|
钱超等: "基于蒙特卡罗模拟的交通状态辨识", 《交通运输系统工程与信息》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109993555A (en) * | 2017-12-30 | 2019-07-09 | 中国移动通信集团四川有限公司 | Internet television potential user complains prediction technique, device and equipment |
CN108564110A (en) * | 2018-03-26 | 2018-09-21 | 上海电力学院 | A kind of Air Quality Forecast method based on clustering algorithm |
CN108564110B (en) * | 2018-03-26 | 2021-07-20 | 上海电力学院 | Air quality prediction method based on clustering algorithm |
CN109213752A (en) * | 2018-08-06 | 2019-01-15 | 国网福建省电力有限公司信息通信分公司 | A kind of data cleansing conversion method based on CIM |
CN109374631B (en) * | 2018-09-30 | 2021-06-29 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | Tunnel state evaluation method |
CN109374631A (en) * | 2018-09-30 | 2019-02-22 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | A kind of tunnel state evaluating method |
WO2021042844A1 (en) * | 2019-09-06 | 2021-03-11 | 平安科技(深圳)有限公司 | Large-scale data clustering method and apparatus, computer device and computer-readable storage medium |
CN110705602A (en) * | 2019-09-06 | 2020-01-17 | 平安科技(深圳)有限公司 | Large-scale data clustering method and device and computer readable storage medium |
CN113447813A (en) * | 2020-09-03 | 2021-09-28 | 鲁能集团有限公司 | Fault diagnosis method and equipment for offshore wind generating set |
CN113447813B (en) * | 2020-09-03 | 2022-09-13 | 中国绿发投资集团有限公司 | Fault diagnosis method and equipment for offshore wind generating set |
CN114821514A (en) * | 2022-06-30 | 2022-07-29 | 交通运输部公路科学研究所 | A remove detecting system for detecting tunnel road surface wet and slippery state |
CN116611620A (en) * | 2023-07-18 | 2023-08-18 | 厚德智能技术(山东)有限公司 | Smart city safety collaborative management information system |
CN116611620B (en) * | 2023-07-18 | 2023-09-19 | 厚德智能技术(山东)有限公司 | Smart city safety collaborative management information system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107451622A (en) | A kind of tunnel operation state division methods based on big data cluster analysis | |
Zhao et al. | Development of a representative urban driving cycle construction methodology for electric vehicles: A case study in Xi’an | |
CN112508392B (en) | Dynamic evaluation method for traffic conflict risk of hidden danger road section of mountain area double-lane highway | |
CN104715605B (en) | A kind of traffic circulation data and emissions data coupling process and system based on VSP distributions | |
Yang et al. | Eco-driving system for connected automated vehicles: Multi-objective trajectory optimization | |
CN104269057B (en) | A kind of bayonet socket sensor deployment method based on Floating Car OD data | |
CN103895649B (en) | A kind of driver safety driving warning method | |
CN106127586A (en) | Vehicle insurance rate aid decision-making system under big data age | |
CN107239844A (en) | A kind of petroleum transportation vehicle oil consumption Forecasting Methodology based on Hadoop | |
CN106682699A (en) | Vehicle exhaust emission characteristic analytic processing method based on clustering analysis | |
CN109086946A (en) | A kind of polluted gas emitted smoke method of conventional energy resource and new energy public transit vehicle | |
CN111142538B (en) | Unmanned ship operation strategy real-time decision-making method for water environment monitoring | |
CN109878509A (en) | Monoblock type tank car Multi-source Information Fusion rollover method for early warning based on fuzzy logic | |
CN106651210A (en) | CAN data-based driver comprehensive quality evaluation method | |
Yang et al. | Identifying intercity freight trip ends of heavy trucks from GPS data | |
CN105912862A (en) | Method for exhaust discharge detection and method and device for atmospheric pollution analysis | |
CN107038860A (en) | A kind of user's driving behavior methods of marking based on ADAS technologies and regression model | |
Zhou et al. | Comparing factors affecting injury severity of passenger car and truck drivers | |
CN104850676A (en) | Highway bridge random traffic flow simulation method | |
CN107180534A (en) | The express highway section average speed method of estimation of support vector regression fusion | |
CN105786895A (en) | Calculating method and device of discharge amount of road intersection | |
CN106781494A (en) | A kind of telemetering motor vehicle tail equipment points distributing method based on track of vehicle Yu flow | |
CN105844915A (en) | Method for determining traffic flow fundamental diagram in variable speed limit control state | |
CN110399364B (en) | Data fusion method based on multiple road detector data | |
CN113095387B (en) | Road risk identification method based on networking vehicle-mounted ADAS |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171208 |