CN111814610B - Ocean platform operation state visualization method based on vibration monitoring - Google Patents
Ocean platform operation state visualization method based on vibration monitoring Download PDFInfo
- Publication number
- CN111814610B CN111814610B CN202010589574.2A CN202010589574A CN111814610B CN 111814610 B CN111814610 B CN 111814610B CN 202010589574 A CN202010589574 A CN 202010589574A CN 111814610 B CN111814610 B CN 111814610B
- Authority
- CN
- China
- Prior art keywords
- data
- color
- characteristic
- platform
- map
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Abstract
The invention relates to a vibration monitoring-based ocean platform operation state visualization method, which comprises the steps of carrying out sensor deployment and control on an ocean platform, then carrying out signal characteristic extraction and carrying out cluster analysis; according to the clustering analysis result, combining the characteristics of the pair of signals to form a digital label describing stationarity and amplitude characteristics; and then, establishing a real-time monitoring characteristic gridding map, and establishing a color-taking criterion and a corresponding criterion of map characteristics and an operation state by grading of the digital tags so as to complete visualization processing.
Description
Technical Field
The invention relates to a visualization method for an operation state of an ocean platform.
Background
The rapid development of society, science and technology and economy can not leave resources such as petroleum, natural gas and the like. As the reserves of onshore oil and gas resources decrease, mankind turns to the ocean to seek more resources. Since the last century, the offshore oil industry has a leap-type development situation, and an important supporting carrier, namely an offshore platform, plays an important role. However, in a severe marine environment, the ocean platform will be subjected to periodic loading by sea wind, sea waves and sea currents for a long time, and also unconventional loading by ship berthing, typhoons, earthquakes and the like. In the process, the platform can be damaged to different degrees, and once the detection is not timely or the decision is misappropriate, the platform is very likely to cause serious economic loss, serious environmental pollution and other consequences.
A structure safety and health assessment method based on vibration signals is a technical means mainly adopted in engineering. However, this kind of evaluation method often lacks monitoring and evaluation on the ocean platform mode, so that the monitoring data directly enters the safety and health evaluation system, causing misjudgment and system abnormality. This is because the data is subjected to an unusual mutation phenomenon due to the change of the operation state during the operation of the platform, which is troublesome for the structural safety evaluation system. Therefore, the safety evaluation of the structure must first form a macroscopic prejudgment on the operating state of the platform, and provide a direction for subsequent safety evaluation. For a complex comprehensive system such as a platform, the mechanical transmission and balance are realized by key component components. Therefore, if the evolution of the platform mode is reasonably judged, monitoring points must be established on key parts of the structure, and a system for monitoring the operation mode is established, so that the directionality and the reliability are provided for subsequent safety and health analysis.
Currently, a monitoring system for an operation mode is often lacked in platform safety and health monitoring based on vibration signals. The method is just aiming at the defect to establish an operation mode monitoring system aiming at the structural characteristics of the ocean platform. Firstly, establishing a sensor grading arrangement control rule aiming at the structural characteristics of the platform, and arranging and controlling sensing equipment at a plurality of positions of the platform to monitor vibration signals. And then, monitoring data are distributed and controlled through the structural characteristics of the analysis platform, data fluctuation characteristics are extracted, and an identification algorithm and a clustering algorithm are established to form a digital label of the fluctuation characteristics of the monitoring points. And finally, realizing visual presentation of the operation mode through the map corresponding criterion.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a vibration monitoring-based ocean platform operation state visualization method by combining the defects of the current vibration signal-based platform health detection technology. The method can rapidly extract the vibration signal characteristics of the platform and visually display the time evolution behavior of the vibration signal characteristics in an image mode. The invention is suitable for processing and analyzing the vibration signal of the ocean platform. The technical scheme of the invention is as follows:
a visual method of an ocean platform operation state based on vibration monitoring comprises the steps of carrying out sensor deployment and control on an ocean platform, then carrying out signal feature extraction and carrying out cluster analysis; according to the clustering analysis result, combining the characteristics of the pair of signals to form a digital label describing stationarity and amplitude characteristics; then, a real-time monitoring feature gridding map is established, a color-taking criterion and a corresponding criterion of map features and operation states are formulated by grading of the digital labels, so that visualization processing is completed, and the method comprises the following aspects:
(1) deploying sensors according to platform characteristics to form a layered sensing network system
The structural vibration response feedback under the environmental load is fed back at the main vertical pile pipe node of the jacket structure according to the structural characteristics of the ocean platform, the vibration feedback of the operating state of the platform is concentrated on the deck positions of each layer of the upper module of the jacket, and the sensors are arranged and controlled according to the following requirements:
1) the sensor distribution and control position is arranged at the intersection point of the main vertical pile of the jacket structure and the deck of the upper module;
2) the sensors are classified according to the structural symmetry and the deck level;
3) the structural strength of the platform cannot be damaged by arrangement or safety problems of components or operation nearby the installation position cannot be caused;
4) ensuring that each sensor does not influence the operation of the platform equipment and simultaneously meeting the space requirement;
(2) digital label processing analysis according to the data feedback signal of control
The method comprises the steps of carrying out time division on platform sensing data, and dividing a data sequence of a sensor into subsequences X i {X 1 ,X 2 …X N N is the number of subsequences, each subsequence maps a time period, the time interval of each subsequence is equal, and for the subsequence X in each time period, the number of subsequences is smaller than the number of subsequences i A three-point extreme method is carried out to obtain data points { p) on all upper envelope lines 1 ,p 2 …p s And s is the number of data points on an envelope line determined by a three-point extreme method, the level and the trend of the integral vibration of the platform are reflected by the data points on the envelope line, and the mean value mu of the data points on the envelope line is selected as the mathematical characteristic of the overall vibration state of the reaction platform:
introducing envelope data standard deviation sigma as a parameter for reflecting the event characteristic of platform vibration, and setting the mean value mu plus one time of standard deviation as a characteristic standard of state identification, namely mu + sigma, wherein sigma is the standard deviation of data points on an envelope line:
data to be storedPoint { p 1 ,p 2 …p s The data of more than the eigenvalue μ + σ in the event are defined as an event characteristic data set { y } 1 ,y 2 …y l Defining eta as the ratio of the number of event characteristic data sets to the number of data characteristic data sets, i.e. defining
For m groups of sensors to obtain data, obtaining a group of corresponding characteristic values, namely calculating a data array X i ＝{(μ 1 ,σ 1 )，(μ 2 ,σ 2 )...(μ m ,σ m ) And calculating the distance between the characteristic values in each group:
a and b are the lower labels of each group of sensors, and the characteristic distances are sorted to obtain a distance extreme value d max ,d min And standardizing all data to obtain a fuzzy similarity matrix:
the transfer closure matrix of the fuzzy similarity matrix is T', and T is calculated 2 、T 4 、...、T 2k Up to T 2k ＝T 2(k-1) Then the transfer closure matrix T ═ T 2k And performing cluster analysis according to the transfer closure matrix: firstly, each group of data is regarded as a class; then, calculating a minimum class interval, and calculating a ambiguity parameter through the minimum class interval; clustering is carried out through the ambiguity parameters to obtain a clustering result, wherein k types are obtained; repeating re-clustering on the obtained clustering result until k is 1, ending clustering calculation to obtain a series of singly-subtracted ambiguity sequences, and taking the ambiguity value with the maximum difference as the optimal clustering ambiguity
Setting the optimal clustering result as X 1 ,X 2 ,…,X k From this, k types of digital labels are formed, with a set X of clusters of n elements for each type i Calculating the center point of the wave-like feature
And a characteristic center point of like amplitude
Wherein n is the number of elements in the clustering geometry; for new vibration signal data, the current vibration signal and each class X need to be calculated i Distance d of i Determining which type of digital label the current vibration signal belongs to, and for m-dimensional data with multiple sensors, classifying the data of each dimension to form an m-dimensional digital label;
(3) establishing a real-time monitoring characteristic grid visualization map
The original data is divided into multiple data segments according to time segments, the divided data segments are regarded as two adjacent small windows without overlapping, and the resolution of each small window is determined as follows
w＝W/N
h＝H/m
Wherein W, H is the width and height of the display resolution, w, h are the width and height of the grid resolution, N is the number of segments, and m is the data dimension; for the formed k-type data labels, carrying out classification RGB color sampling according to fluctuation characteristics and amplitude characteristics;
for data segment x, the fluctuation characteristic σ is calculated i And assigned feature mu i To realize RGB color sampling, the following normalization process is performed:
translation of fluctuation characteristics for each class of tags
Wherein
The standard fluctuation digital label is in a range of 0-1, and the logarithmic base a is larger than 1;
to highlight its amplitude characteristics, the amplitude characteristics are also normalized logarithmically:
wherein
The range of the normalized amplitude digital label is between 0 and 1, the logarithm base b is less than 1, and the data color characteristics are formed by the method
And
when in use
When the color is in a color of the primary R color, the fluctuation characteristic is mainly indicated, and the primary R color is the main control color
Color taking is carried out; at this time when
The higher the color, the closer the color is to red, and
because the blue degree of the color can be properly adjusted, alpha is between 0 and 1;
when in use
When the color is in a primary color, the amplitude characteristic is used as the main color, and the primary color B is used as the main control color
Color taking is carried out; at this time when
The higher the color, the closer the color is to blue, and
then adjust the red degree of the block properly, beta is between 0 and 1.
(4) Establishing a corresponding criterion of the map characteristics and the operation state: dividing the operation state into different kinds of operation states, and establishing a corresponding criterion of the map characteristics and the operation state according to the determined corresponding relation, wherein the method comprises the following steps:
for m-dimensional data in a certain time period, calculating the overall fluctuation map characteristic lambda:
and overall amplitude map features θ:
and determining the corresponding relation and the corresponding criterion of the map characteristics and different types of operation states according to the value ranges of the lambda and the theta.
Drawings
FIG. 1 is a diagram of a particular example of the present invention of a method for controlling ocean platform sensors in a hierarchical manner, (a) is a schematic diagram of the main column and deck intersection points named according to the elevation (A, B) position, wherein A1-A9 and B1-B9 represent A, B elevation main column and deck intersection points, respectively. (b) Is a schematic diagram of the main upright post and deck intersection points named according to the deck positions, wherein U1-U6, M1-M6 and L1-L6 represent the intersection points of the upper deck, middle deck and lower deck with the main upright post respectively.
FIG. 2 is a table acceleration signal display diagram according to an embodiment of the present invention
FIG. 3 is a visualization atlas grid map in an embodiment of the invention
FIG. 4 shows the stage states, (a) under low load operation and (b) under high load operation in an embodiment of the present invention
FIG. 5 shows the platform status in an embodiment of the present invention, (a) is the storm maintenance operation status (confirmed by the platform side that the site has a Liqima typhoon crossing the scene at the present time), (b) is the special construction operation status (confirmed by the platform that the rust removal and painting operation is performed at the present time)
The reference numbers in the figures illustrate: in fig. 1, a1-a9 represent intersection points of a riser main column and a deck of a jacket structure a, B1-B9 represent intersection points of a riser main column and a deck of a jacket structure B, U1-U6 represent intersection points of an upper deck and a main column of a jacket upper module, M1-M6 represent intersection points of a middle deck and a main column of a jacket upper module, and L1-L6 represent intersection points of a lower deck and a main column of a jacket upper module.
Detailed Description
According to the fluctuation characteristics of the acceleration signals of the ocean platform, the following technical scheme is adopted to carry out sensor deployment and control on the platform, and then signal characteristic extraction and cluster analysis are carried out. And according to the clustering analysis result, combining the characteristics of the signals to form a digital label which is carefully described on stationarity and amplitude. And then establishing a real-time monitoring characteristic gridding map, and establishing a color-taking criterion and a corresponding criterion of map characteristics and an operation state by the grading of the digital tags so as to complete visualization processing. Finally, the corresponding criterion or the judgment rule of the map characteristic and the operation state is realized. The specific implementation method comprises the following steps:
1. firstly, sensors are distributed and controlled according to platform characteristics to form a layered sensing network system. The vibration sensor is mainly controlled by capturing the structural vibration response of environmental load and monitoring the vibration response caused by special operation. Structural vibration response under environmental load is usually fed back at the main riser pipe node of the jacket structure, and vibration feedback of the working state of the platform is mainly concentrated on the deck position of each layer of the upper module of the jacket. According to the structural characteristics of the ocean platform, the overall deployment and control principle of the sensor is established as follows:
1) the sensor deployment and control positions are arranged at the intersection point of the main vertical pile of the jacket structure and the deck of the upper module.
2) The sensors are classified according to structural symmetry and deck level (taking a three-deck structure as an example, as shown in fig. 1), and main deployment points can be deployed according to the elevation of a structure A, B (the elevation A1-A9 and the elevation B1-B9 of a jacket structure in fig. 1) or the deployment points can be deployed according to the principle of the deck level (the upper module upper deck U1-U6, the middle deck M1-M6 and the lower deck L1-L6 of a jacket in fig. 1).
3) The control must not damage the structural strength of the platform itself or pose safety issues to components or operations near the installation site.
4) And the operation of the platform equipment is not influenced by each sensor, and the space requirement is met.
On the premise of meeting the overall deployment and control principle, the monitoring of vibration response caused by special operation is met, so that the following facilities or modules need to be approached:
1) an upper deck: drilling (DSM) module, crane and living building
2) Middle deck: central control room, diesel generator, air conditioner, water pump and other equipment
3) The lower deck: apparatus such as riser area, emergency machine and separator
Taking a certain three-layer deck platform as an example, the sensor control is carried out on the three-layer deck according to the control principle, and 18 structural vibration test sensors are used in total. The positions of the counterweight arrangement and control measuring points are illustrated, and the arrangement and control measuring points in the living area can monitor the activities of platform personnel and the lifting and landing of the helicopter. The monitoring of the crane activity can be realized to the cloth accuse measurement station of platform crane. The control of well moving and drilling operation can be realized through the control points of the DSM module. The monitoring of the activity of the generator can be realized by the distribution and control measuring point beside the generator.
2. And carrying out digital label processing analysis according to the controlled data feedback signal. And carrying out time segmentation on the platform sensing data. Segmenting a data sequence of a sensor into subsequences X i {X 1 ,X 2 …X N And N is the number of subsequences. Each sub-sequence is mapped to a time segment and the time intervals of each sub-sequence are equal. For the sub-sequence X in each time segment i A three-point extreme method is carried out to obtain data points { p) on all upper envelope lines 1 ,p 2 …p s And s is the number of data points on all the envelope lines determined by a three-point extreme method. Because of the particularity of the environmental condition of the platform, the level and the trend of the overall vibration of the platform can be reflected by the data points on the envelope line, and therefore the average value mu of the data points on the envelope line is selected as the mathematical characteristic reflecting the overall vibration state of the platform, namely
On the other hand, for various special states of the ocean platform, such as common construction, well drilling and the like, the data are inevitably subjected to fluctuation caused by state change on the basis of the original fluctuation. Since such fluctuations are often hardly reflected in the overall fluctuation feature value, the envelope data standard deviation σ is introduced as a parameter reflecting the event feature of the stage vibration, and the mean value plus one time standard deviation is set as the feature standard of state recognition, i.e., μ + σ, where σ is the standard deviation of data points on the envelope line, and is calculated by the formula of
Envelope data { p } 1 ,p 2 …p s The data of more than the eigenvalue μ + σ in the event are defined as an event characteristic data set { y } 1 ,y 2 …y l Defining eta as the number and data of event feature data setsRatio of number of characteristic data, i.e.
For m groups of sensors to obtain data, a group of corresponding characteristic values can be obtained, and then the data array X can be calculated i ＝{(μ 1 ,σ 1 )，(μ 2 ,σ 2 )...(μ m ,σ m ) Calculating the distance between the characteristic values in each group
Wherein a and b are the lower labels of each group of sensors, and the distance extreme value d can be obtained by sorting the characteristic distances max ,d min And standardizing all the data to obtain a fuzzy similar matrix
Then the transitive closure matrix of the fuzzy similarity matrix is T', and T is calculated 2 、T 4 、...、T 2k Up to T 2k ＝T 2(k-1) Then the transfer closure matrix T ═ T 2k . And (3) carrying out clustering analysis according to the transfer closure matrix:
1) let y a ,y b E X, where a, b1, 2.
2) Calculating the distance between the minimum classes:
and obtaining an ambiguity parameter alpha t ：
3) Will pass closure matrix satisfying T' ab ≥α t Conditions ofEnvelope data p of (1) a And p b Classifying into one class to obtain clustering result X 1 ,X 2 ,…,X q Where q is the ambiguity parameter α t And (5) clustering to obtain the target.
4) If k is>1, then let y i ＝X i ,y j ＝X j I, j ═ 1,2, …, k, equations (2) and (3) are repeatedly calculated; if k is 1, the next step is performed.
5) Calculating the difference of the fuzzy degree: delta alpha i ＝α i -α i+1 1,2, …, t-1, and z is argmin Δ α i Then give an order
6) Will pass the satisfaction in the closure matrix
Envelope data p of the condition i And p j Classifying into one class to obtain clustering result X 1 ,X 2 ,…,X k . And obtaining the optimal clustering result of the data.
The clustering algorithm is to regard each group of data as one type; then, calculating a minimum class interval, and calculating a ambiguity parameter through the minimum class interval; clustering is carried out through the ambiguity parameters to obtain clustering results, wherein the clustering results are k types; and repeatedly re-clustering the obtained clustering result (calculating the minimum class distance, calculating the ambiguity parameter through the minimum class distance, and clustering through the ambiguity parameter) until k is 1. And obtaining a series of singly-subtracted ambiguity sequences, and taking the ambiguity value at the position with the maximum difference (namely the position with the maximum minimum class spacing difference) as the optimal clustering ambiguity. And clustering by using the ambiguity to obtain a k-class result, namely the classification criterion of the platform vibration level.
For forming clustering result X 1 ,X 2 ,…,X k . A class k digital label may be formed. Set of clusters X with n elements for each class i Calculating the center point of the wave-like feature
And a magnitude-like feature center point
Wherein n is the number of elements in the cluster geometry. Thus, for new vibration signal data X, X needs to be calculated and each class X i Distance d of i To determine which type of tag should be marked on the current data signal x, i.e.
Wherein, mu i x Mean value, σ, of vibration signal data x i x Is the standard deviation of the vibration signal data x. Then for any data x, the minimum distance min d should be chosen 1 ,d 2 ...d k The corresponding category is used as the data number label. For m-dimensional data x with multiple sensors, the data of each dimension in x needs to be classified to form an m-dimensional digital label.
3. And establishing a real-time monitoring characteristic grid visualization map. Since the original data is split into multiple segments, the split data segments can be viewed as two adjacent small windows that do not overlap. The resolution of each small window is determined as follows
Where W, H is the width and height of the display resolution, w, h are the width and height of the grid resolution, N is the number of segments, and m is the data dimension. This is sufficient to characterize the data while ensuring reasonably low storage requirements. On the other hand, for the formed k-type grid label, the color sampling of the graded RGB colors is required according to the fluctuation characteristic and the amplitude characteristic.
For data segment x, the fluctuation characteristic σ is calculated i And an assigned feature mu i . For standard RGB color sampling, a normalization process is required. It is noted that the level of the fluctuation characteristics is often influenced to a great extent by the changes of noise, construction and platform states, and if the fluctuation characteristics of a certain type of labels are too high, the dispersion of the fluctuation characteristics of other types of labels is often lost. The fluctuation characteristics for each type of label need to be converted into
Wherein
The standard fluctuation digital label is in a range of 0-1. The choice of the logarithmic base a is generally greater than 1. Also, since changes in the state of the platform are less likely to cause changes in the amplitude, to highlight its amplitude characteristics, the amplitude characteristics are also normalized logarithmically:
wherein
The range of the normalized amplitude digital label is 0-1. The choice of the logarithmic base b is generally less than 1. By the method, the color characteristics of the data can be formed
And
when in use
When the color is in the middle of the color, the fluctuation characteristic is mainly used, and the main R color is used as the main controlColor button
And (5) carrying out color extraction. At this time when
The higher the color, the closer the color is to red, and
since the blue color degree of the color is properly adjusted, α is generally between 0 and 1.
When in use
When the color is in a primary color, the amplitude characteristic is used as the main color, and the primary color B is used as the main control color
And (5) carrying out color extraction. At this time when
The higher the color, the closer the color is to blue, and
the red level of the block is properly adjusted, and β is usually between 0 and 1.
4. And establishing corresponding criteria of the map characteristics and the operation state. Establishing a corresponding criterion of the map characteristic and the operation state according to the determined corresponding relation according to the well-regulated operation state (such as a high-load production operation state, a medium-load production operation state, a low-load production operation state, a storm maintenance production operation state, a well workover operation state and a special construction operation state). The method is as follows
For data x with m dimensions in a certain time period, the atlas characteristic value is the atlas characteristic set calculated in step 3
With features of global wave pattern lambda
And overall amplitude map feature θ
The corresponding relation between the map characteristics and the operation state can be further determined according to the value ranges of the lambda and the theta. The following is a description of a specific example. And processing the vibration acceleration data of a certain platform, wherein the vibration acceleration sensors of the platform have 18 groups in total, and the time interval is set to be 1 hour. The acceleration signal is shown in fig. 2, and the classification result k is calculated to be 12 types according to the steps 2 and 3 of the present invention. The 12 types of digital labels form the stable map features and the vibration map features according to the step 3, the RGB colors are given to the stable map features and the vibration map features, and the classification results are obtained according to the time sequence as shown in FIG. 3. From fig. 3, the whole process of the change of the vibration level of the platform with time can be observed clearly, and then through step 4, the corresponding relation between the map characteristics and the operation state of the platform is determined, as shown in fig. 4 to 5. Wherein, fig. 4(a) is a low-load operation state, fig. 4(b) is a high-load operation state, fig. 5(a) is a storm maintenance production operation state (confirmed by a platform side, a site at the present stage has a lee typhoon crossing), and fig. 5(b) is a special construction operation state (confirmed by the platform that a rust removing and painting operation is carried out at the present stage).
Claims (1)
1. A visual method of an ocean platform operation state based on vibration monitoring comprises the steps of carrying out sensor deployment and control on an ocean platform, then carrying out signal feature extraction and carrying out cluster analysis; forming a digital label describing stationarity and amplitude characteristics by combining the characteristics of the pair signals according to the clustering analysis result; then, a real-time monitoring feature gridding map is established, and a color taking criterion and a corresponding criterion of map features and an operation state are formulated by the grading of the digital tags, so that the visualization processing is completed; the method comprises the following aspects: (1) deploying sensors according to platform characteristics to form a layered sensing network system
The structural vibration response feedback under the environmental load is fed back at the main vertical pile pipe node of the jacket structure according to the structural characteristics of the ocean platform, the vibration feedback of the operating state of the platform is concentrated on the deck positions of each layer of the upper module of the jacket, and the sensors are arranged and controlled according to the following requirements:
1) the sensor distribution and control position is arranged at the intersection point of the main vertical pile of the jacket structure and the deck of the upper module;
2) the sensors are classified according to the structural symmetry and the deck level;
3) the structural strength of the platform cannot be damaged by arrangement or safety problems of components or operation nearby the installation position cannot be caused;
4) ensuring that each sensor does not influence the operation of the platform equipment and simultaneously meeting the space requirement;
(2) digital label processing analysis according to distributed control data feedback signal
The method comprises the steps of carrying out time division on platform sensing data, and dividing a data sequence of a sensor into subsequences X i :{X 1 ,X 2 …X N N is the number of subsequences, each subsequence maps a time period, the time interval of each subsequence is equal, and the subsequence X in each time period is subjected to multiple-stage mapping i A three-point extreme method is carried out to obtain all envelope characteristic data points { p } 1 ,p 2 …p s S is the number of data points on an envelope line determined by a three-point extreme method, the level and the trend of the overall vibration of the platform are reflected by the data points on the envelope line, and the mean value mu of the characteristic data points on the envelope line is selected as the mean value characteristic of the overall vibration state of the reaction platform:
introducing an envelope data standard deviation sigma as a parameter reflecting the event characteristic of the platform vibration, and setting a mean value mu plus one time of the standard deviation as a characteristic standard of state identification, namely mu + sigma, wherein sigma is the standard deviation of data points on an envelope:
envelope data points { p } 1 ,p 2 …p s The data of more than the eigenvalue μ + σ in the event are defined as an event characteristic data set { y } 1 ,y 2 …y l Where l is the number of event feature data sets, the numerical label η is defined as the ratio of the number of event feature data sets to the number of feature data sets, i.e.
For m groups of sensors to obtain data, obtaining a group of corresponding characteristic values, i.e. calculating a data array
Further, the numerical label { η [ ] can be calculated according to equation (3) 1 ,η 2 …η m Then the characteristic distance d obtained from each set of characteristic values can be calculated ab ：
A and b are the lower labels of each group of sensors, and the characteristic distances are sorted to obtain a distance extreme value d max ,d min And standardizing all data to obtain a fuzzy similarity matrix:
the transfer closure matrix of the fuzzy similarity matrix is T', and T is calculated 2 、T 4 、...、T 2k Up to T 2k ＝T 2(k-1) Then the transfer closure matrix T ═ T 2k And performing cluster analysis according to the transfer closure matrix: firstly, each group of data is regarded as a class; then, calculating a minimum class interval, and calculating a ambiguity parameter through the minimum class interval; clustering is carried out through the ambiguity parameters to obtain clustering results, wherein the clustering results are k types; repeating re-clustering on the obtained clustering result until k is 1, ending clustering calculation to obtain a series of singly-subtracted ambiguity sequences, and taking the ambiguity value with the maximum difference as the optimal clustering ambiguity
Setting the optimal clustering result as X 1 ,X 2 ,…,X k From this, k classes of digital labels are formed, X for each class of a set of clusters with n elements i Calculating the center point of the wave-like feature
And a magnitude-like feature center point
Wherein n is the number of elements in the cluster set; for new vibration signal data, the current vibration signal and each class X need to be calculated i Distance d of i Determining which type of digital label the current vibration signal belongs to, and for m-dimensional data with multiple sensors, classifying the data of each dimension to form an m-dimensional digital label;
(3) establishing a real-time monitoring characteristic grid visualization map
The original data is divided into multiple data segments according to time segments, the divided data segments are regarded as two adjacent small windows without overlapping, and the resolution of each small window is determined as follows
Wherein W, H is the width and height of the display resolution, w, h are the width and height of the grid resolution, N is the number of segments, and m is the data dimension; for the formed k-type data labels, carrying out classification RGB color sampling according to fluctuation characteristics and amplitude characteristics;
for data segment x, the fluctuation characteristic σ is calculated i And an assigned feature mu i To realize RGB color sampling, the following normalization process is performed:
translation of fluctuation characteristics for each class of tags
Wherein
The standard fluctuation digital label is a standardized fluctuation digital label, the range is 0-1, and the logarithmic base a is larger than 1; to highlight its amplitude characteristics, the amplitude characteristics are also normalized logarithmically:
wherein
The range of the normalized amplitude digital label is between 0 and 1, the logarithm base b is less than 1, and the data color characteristics are formed by the method
And
when in use
When the color is in a color of the primary R color, the fluctuation characteristic is mainly indicated, and the primary R color is the main control color
Color taking is carried out; at this time when
The higher the color, the closer the color is to red, and
because the blue degree of the color can be properly adjusted, alpha is between 0 and 1;
when in use
When the color is in a primary color, the amplitude characteristic is used as the main color, and the primary color B is used as the main control color
Color taking is carried out; at this time when
The higher the color, the closer the color is to blue, and
properly adjusting the red degree of the block, wherein beta is between 0 and 1;
(4) establishing a corresponding criterion of the map characteristics and the operation state: dividing the operation state into different kinds of operation states, and establishing a corresponding criterion of the map characteristics and the operation state according to the determined corresponding relation, wherein the method comprises the following steps:
for m-dimensional data in a certain time period, calculating the overall fluctuation map characteristic lambda:
and overall amplitude map features θ:
according to the value ranges of the lambda and the theta, the corresponding relation and the corresponding criterion of the map characteristics and different types of operation states can be determined.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010589574.2A CN111814610B (en) | 2020-06-24 | 2020-06-24 | Ocean platform operation state visualization method based on vibration monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010589574.2A CN111814610B (en) | 2020-06-24 | 2020-06-24 | Ocean platform operation state visualization method based on vibration monitoring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111814610A CN111814610A (en) | 2020-10-23 |
| CN111814610B true CN111814610B (en) | 2022-09-06 |
Family
ID=72856495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010589574.2A Active CN111814610B (en) | 2020-06-24 | 2020-06-24 | Ocean platform operation state visualization method based on vibration monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111814610B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102855662A (en) * | 2012-07-25 | 2013-01-02 | 中国科学院对地观测与数字地球科学中心 | Ocean environment visualization method |
| CN103398860A (en) * | 2013-07-26 | 2013-11-20 | 天津亿利科能源科技发展股份有限公司 | Method for monitoring safety of ocean platform on basis of displacement sensors |
| WO2013180496A2 (en) * | 2012-05-30 | 2013-12-05 | 주식회사 싸이트로닉 | System and method for providing information on fuel savings, safe operation, and maintenance by real-time predictive monitoring and predictive controlling of aerodynamic and hydrodynamic environmental internal/external forces, hull stresses, motion with six degrees of freedom, and the location of marine structure |
| CN110081930A (en) * | 2019-04-24 | 2019-08-02 | 中国科学院城市环境研究所 | A kind of land, sea and air integration ecological environmental monitoring system |
| CN110110352A (en) * | 2019-03-20 | 2019-08-09 | 浙江海洋大学 | A kind of ship longitudinal direction boat state and Resistance Prediction method based on overlapping grid |
-
2020
- 2020-06-24 CN CN202010589574.2A patent/CN111814610B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013180496A2 (en) * | 2012-05-30 | 2013-12-05 | 주식회사 싸이트로닉 | System and method for providing information on fuel savings, safe operation, and maintenance by real-time predictive monitoring and predictive controlling of aerodynamic and hydrodynamic environmental internal/external forces, hull stresses, motion with six degrees of freedom, and the location of marine structure |
| CN102855662A (en) * | 2012-07-25 | 2013-01-02 | 中国科学院对地观测与数字地球科学中心 | Ocean environment visualization method |
| CN103398860A (en) * | 2013-07-26 | 2013-11-20 | 天津亿利科能源科技发展股份有限公司 | Method for monitoring safety of ocean platform on basis of displacement sensors |
| CN110110352A (en) * | 2019-03-20 | 2019-08-09 | 浙江海洋大学 | A kind of ship longitudinal direction boat state and Resistance Prediction method based on overlapping grid |
| CN110081930A (en) * | 2019-04-24 | 2019-08-02 | 中国科学院城市环境研究所 | A kind of land, sea and air integration ecological environmental monitoring system |
Non-Patent Citations (3)
| Title |
|---|
| ""Nonlinear dynamic analysis and fatigue damage assessment for a deepwater test string subjected to random loads";Kang Liu等;《Pet. Sci. (2016)》;20151201;第126-134页 * |
| "I4Ocean: An Interactive Simulation and Scientific Visualization Platform for Marine Application";Pengbo Ji等;《Rev. Téc. Ing. Univ. Zulia》;20161231;第39卷(第5期);第169-179页 * |
| "面向海洋平台的无线低功耗冰激振动监测系统";慕昆等;《现代电子技术》;20180101;第41卷(第1期);第112-116页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111814610A (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022077605A1 (en) | Wind turbine blade image-based damage detection and localization method | |
| CN112199993B (en) | Method for identifying transformer substation insulator infrared image detection model in any direction based on artificial intelligence | |
| CN109871976A (en) | A kind of prediction technique of power quality containing distributed power distribution network based on cluster and neural network | |
| CN110985310B (en) | Wind driven generator blade fault monitoring method and device based on acoustic sensor array | |
| CN110533331B (en) | Fault early warning method and system based on transmission line data mining | |
| CN110070530A (en) | A kind of powerline ice-covering detection method based on deep neural network | |
| CN110196962A (en) | A kind of air speed abnormality recognition method based on Density Estimator | |
| CN114856941A (en) | Offshore wind power plant and unit fault diagnosis operation and maintenance system and diagnosis operation and maintenance method thereof | |
| CN110555478B (en) | Fan multi-fault diagnosis method based on depth measurement network of difficult sample mining | |
| CN115099335A (en) | Abnormal identification and feature screening method and system for multi-source heterogeneous data | |
| CN111814600B (en) | Ocean platform operation mode monitoring system | |
| CN112464439A (en) | Three-layer data mining-based power system fault type analysis method and device | |
| CN113569459A (en) | Group intelligence-based hydraulic support performance degradation quantitative evaluation method and system | |
| CN111814610B (en) | Ocean platform operation state visualization method based on vibration monitoring | |
| CN109902265B (en) | Underground early warning method based on hidden Markov model | |
| CN112802011A (en) | Fan blade defect detection method based on VGG-BLS | |
| CN117556197A (en) | Typhoon vortex initialization method based on artificial intelligence | |
| CN115202404B (en) | Maintenance and inspection method for photovoltaic power generation set based on unmanned aerial vehicle | |
| CN113606099A (en) | Method and system for monitoring icing of blades of wind turbine generator | |
| CN113538971A (en) | Risk ship detection method based on multi-source heterogeneous data fusion | |
| LU503778B1 (en) | The method for wind speed downscaling of wind farms based on super-resolution reconstruction technology | |
| CN110944354B (en) | Base station interference monitoring method and system based on waveform analysis and deep learning | |
| CN116032003A (en) | Power grid inspection method and system based on power private network data processing | |
| CN116597693A (en) | Inland navigation monitoring system and method | |
| CN115826454A (en) | Large ship vibration isolation system, edge calculation system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |