CN115618255A - Working condition division method, device, equipment and medium - Google Patents
Working condition division method, device, equipment and medium Download PDFInfo
- Publication number
- CN115618255A CN115618255A CN202211329069.XA CN202211329069A CN115618255A CN 115618255 A CN115618255 A CN 115618255A CN 202211329069 A CN202211329069 A CN 202211329069A CN 115618255 A CN115618255 A CN 115618255A
- Authority
- CN
- China
- Prior art keywords
- working condition
- point
- determining
- division
- value
- 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
- 238000000034 method Methods 0.000 title claims abstract description 98
- 238000004458 analytical method Methods 0.000 claims abstract description 60
- 230000008569 process Effects 0.000 claims abstract description 11
- 238000004590 computer program Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 12
- 238000000638 solvent extraction Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000007621 cluster analysis Methods 0.000 description 22
- 238000004422 calculation algorithm Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 238000007781 pre-processing Methods 0.000 description 7
- 239000013598 vector Substances 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000003064 k means clustering Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000013450 outlier detection Methods 0.000 description 3
- 238000010977 unit operation Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a working condition division method, a working condition division device, equipment and a medium, which are applied to the field of process industry. The working condition division method comprises the steps of firstly obtaining each working condition parameter in working condition parameter historical data, carrying out one-dimensional clustering analysis on each working condition parameter, combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis on each working condition parameter into a working condition interval, and determining division of real-time working condition parameters corresponding to working conditions according to the working condition interval. The application also provides a working condition division device, equipment and a medium, which correspond to the working condition division method, so that the beneficial effects are achieved.
Description
Technical Field
The present application relates to the field of process industry, and in particular, to a method, an apparatus, a device, and a medium for dividing operating conditions.
Background
The thermal power generating unit is influenced by the operating mode, and the unit moves under the external restraint operating mode condition of difference, and efficiency produces the difference, because of the power station unit generally faces external environment and the big scheduling problem of unit load change, the unit operating mode change is great to it is also great to lead to thermal power generating unit operating efficiency difference, for guaranteeing thermal power generating unit's operating efficiency, need divide the operating mode condition, selects the optimum external operating mode condition of thermal power generating unit operation. The working condition division method mainly comprises an equal width method, an equal frequency method and a clustering analysis method.
At present, a K-means cluster analysis method is used for dividing the working conditions of the thermal power plant unit, after data preprocessing is carried out on working condition parameters, cluster analysis is carried out, the confirmed optimal cluster number is manually specified, and a working condition result is finally obtained.
Therefore, a new working condition division method is sought, which is a technical problem to be solved urgently by the technical personnel in the field.
Disclosure of Invention
The application aims to provide a working condition division method, a working condition division device, a working condition division equipment and a working condition division medium, wherein a one-dimensional K-means + + clustering analysis algorithm is respectively carried out on each working condition parameter to generate a cluster interval, and then generated cluster results are combined to form the working condition interval, so that the condition that the non-spherical cluster and high-dimensional feature vector clustering effect of the K-means clustering analysis method is not obvious can be avoided.
In order to solve the technical problem, the present application provides a method for dividing operating conditions, including:
acquiring each working condition parameter in the working condition parameter historical data;
combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval;
and determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
Preferably, after acquiring each operating condition parameter in the operating condition parameter history data, the method further comprises:
removing bad point data according to the quality level of the working condition parameter historical data;
eliminating out-of-limit data according to preset upper and lower limits of working condition parameters;
and judging the stability of the working condition parameters according to the stability criterion, and rejecting unstable working condition parameter values in the working condition parameters.
Preferably, the determining the stability of the operating condition parameter according to the stability criterion includes:
judging whether the working condition parameters meet the condition that the difference between the maximum value and the minimum value in the data sequence of the working condition parameters is smaller than a preset value, and the quotient of the process deviation amount and the standard deviation of the data sequence which is 3 times larger than a specified value;
if so, determining that the working condition parameters are stable;
if not, determining that unstable working condition parameter values exist in the working condition parameters, and entering a step of rejecting the unstable working condition parameter values in the working condition parameters.
Preferably, after the stability of the working condition parameters is judged according to the stability criterion and unstable working condition parameter values in the working condition parameters are removed, the method further comprises the following steps:
determining the kth reachable distance of each point in the kth distance neighborhood of each point;
determining a local kth local reachable density of each point;
determining a kth local outlier factor for each point;
and determining the data points to which the maximum n local outlier factors belong, and outputting an outlier set.
Preferably, the performing the K-means + + clustering analysis on the operating condition parameters includes:
determining an initialization clustering center;
determining the distance between each point and the central point;
determining the average value of the sample data in each cluster as a new cluster center;
judging whether the distance between the new clustering center and the old clustering center is smaller than a threshold value;
if yes, generating a central point and each cluster set;
if not, returning to the step of calculating the distance between each point and the central point.
Preferably, determining the initialized cluster center comprises:
randomly selecting a point as a first central point;
determining the distance between each point and the central point;
determining a probability value for each distance;
accumulating the probability value of each point according to a wheel disc method to generate a probability interval;
randomly generating a numerical value within a preset range, and taking a maximum value corresponding point of an interval where the numerical value is located as a next central point;
judging whether the number of the central points is greater than k;
if not, returning to the step of calculating the distance between each point and the central point.
Preferably, before combining the clustering analysis results obtained after performing one-dimensional K-means + + clustering analysis on each operating condition parameter into an operating condition interval, the method further includes:
traversing k;
performing K-means + + clustering analysis with the cluster number equal to K to generate clustering results and K central points;
calculating the contour coefficient of the current k to generate a plurality of contour coefficients;
sequencing the contour coefficients, and determining k corresponding to the maximum value as a recommended k value;
storing the clustering analysis result of the working condition parameters corresponding to the recommended k value in a working condition interval division configuration table for unit optimization so as to provide a division rule for the working condition division of the real-time working condition parameters;
wherein, the range of k is a preset initial value and a maximum value, and if the maximum value is not preset, the maximum value is determined to be
n is the number of samples.
In order to solve the technical problem, the present application further provides a device is divided to operating mode, including:
the acquisition module is used for acquiring each working condition parameter in the working condition parameter historical data;
the combination module is used for combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval;
and the division module is used for determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
In order to solve the technical problem, the present application further provides a working condition division apparatus, including a memory for storing a computer program;
and the processor is used for realizing the steps of the working condition division method when executing the computer program.
In order to solve the above technical problem, the present application further provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the above operating condition division method.
The working condition division method comprises the steps of firstly obtaining each working condition parameter in working condition parameter historical data, carrying out one-dimensional clustering analysis on each working condition parameter, combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis on each working condition parameter into a working condition interval, and determining division of real-time working condition parameters corresponding to working conditions according to the working condition interval.
The application also provides a working condition division device, equipment and a medium, which correspond to the working condition division method, so that the working condition division method has the same beneficial effects as the working condition division method.
Drawings
In order to more clearly illustrate the embodiments of the present application, the drawings required for the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained by those skilled in the art without inventive effort.
Fig. 1 is a flowchart of a working condition division method provided in an embodiment of the present application;
FIG. 2 is a general schematic diagram of a working condition division method according to an embodiment of the present disclosure;
FIG. 3 is a combined flowchart of a method for dividing operating conditions according to an embodiment of the present disclosure;
FIG. 4 is a flowchart of a method for partitioning operating conditions according to another embodiment of the present disclosure;
fig. 5 is a schematic diagram of a thermal power operating condition division method according to another embodiment of the present application;
FIG. 6 is a flow chart of a K-Means + + flow chart provided in accordance with another embodiment of the present application;
FIG. 7 is a flowchart of the initialization center point of the K-Means + + algorithm according to another embodiment of the present application;
FIG. 8 is a flow chart of determining a number of cluster analysis clusters according to another embodiment of the present application;
fig. 9 is a structural diagram of a working condition division apparatus according to another embodiment of the present application;
fig. 10 is a structural diagram of a working condition division apparatus according to another embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without any creative effort belong to the protection scope of the present application.
The power station unit generally faces the problems of large change of the external environment temperature and the unit load and the like, so that the unit operation condition change is large, the characteristic difference of the thermal power unit under different operation conditions is large, and the corresponding optimal values are different. In actual power production process, because of receiving the influence of external environment temperature, unit load change and other factors, lead to the unable state that maintains high-efficient operation for a long time of thermal power unit, the unit moves under the external constraint operating mode condition of difference, and the efficiency of boiler can have very big difference, so, seek the optimal external operating mode condition of unit operation in order to improve the problem that unit operating efficiency needs the solution at present urgently.
The current working condition division method mainly comprises an equal width method, namely, division is carried out according to the boundary condition variation range of the unit, and all working conditions of the unit can be covered; the constant frequency method is that the working condition parameters are divided into k intervals according to the same sample data in each interval; and (4) a clustering analysis method, namely performing working condition division by using a K-means or FCM clustering algorithm generally. The K-means cluster analysis method is most widely applied in practical application, after data preprocessing is carried out on working condition parameters, cluster analysis is carried out, the confirmed optimal cluster number is manually specified, and a working condition result is finally obtained.
The core of the application is to provide a working condition division method, a working condition division device, equipment and a medium, and the problem that the clustering effect of non-spherical clusters and high-dimensional feature vectors is not obvious can be solved.
It should be noted that the working condition division method in the present application is a working condition division method for a thermal power generating unit, and in addition, the working condition division method mentioned in the present application may be implemented by a Micro Control Unit (MCU) or other types of control devices, which does not affect the implementation of the present technical method.
In order that those skilled in the art will better understand the disclosure, the following detailed description will be given with reference to the accompanying drawings.
The application provides a working condition division method, and fig. 1 is a flowchart of the working condition division method provided by the embodiment of the application; as shown in fig. 1, the method includes:
s10: and acquiring each working condition parameter in the working condition parameter historical data.
In specific implementation, the operating condition parameter history data is sequentially imported to obtain each operating condition parameter, and it is understood that the obtaining operation in this embodiment may be completed by the MCU or by other types of controllers, and each operating condition parameter in the operating condition parameter history data may be obtained in real time, that is, after the MCU or other types of controllers obtain each operating condition parameter in the operating condition parameter history data, the data is uploaded in real time, or may be obtained at regular time, that is, after the MCU or other types of controllers obtain each operating condition parameter in the operating condition parameter history data in one period, the data is uploaded according to a preset period.
It should be noted that, the specific type of the operating condition parameter in this embodiment is not limited, and may be an ambient temperature parameter, a unit load parameter, or the like.
S11: and combining the clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval.
In specific implementation, one-dimensional K-means + + clustering analysis is carried out on each working condition parameter, and compared with a K-means clustering analysis method used for dividing the current working conditions, the K-means + + clustering method does not need to artificially determine an initialized clustering center, the number K of the clustering centers does not need to be given in advance, and the initialized clustering center is random. FIG. 2 is a general schematic diagram of a working condition division method according to an embodiment of the present disclosure; as shown in fig. 2, the historical data of the operating condition parameters are sequentially imported, the imported operating condition parameters can be preprocessed and outlier screening can be performed, firstly, elimination of dead points and over-limit points in a data sequence can be completed through data preprocessing operation, and unstable operating condition parameter values are eliminated by judging the stability of the data according to parameter stability criteria. Thirdly, local outlier factor detection processing is carried out on the preprocessed data, the problem of center deviation caused by outliers can be effectively avoided, then a one-dimensional K-means + + clustering analysis algorithm is carried out on each working condition parameter to generate a clustering range interval, each working condition parameter obtains a clustering analysis result, and finally the clustering analysis results of each working condition parameter are combined to form a working condition interval.
S12: and determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
In specific implementation, after one-dimensional K-means + + clustering analysis is performed on each working condition parameter, clustering analysis results after clustering are combined into working condition intervals, so that the working condition intervals are formed based on historical data, the interval parameters are provided for real-time data to divide the working conditions, and as shown in FIG. 2, data support is provided for subsequent real-time data optimization.
FIG. 3 is a combined flowchart of a method for dividing operating conditions according to an embodiment of the present disclosure; as shown in fig. 3, in the working condition division method provided by the present application, after the historical working condition parameters are preprocessed, the retained data is subjected to local outlier detection, which can effectively solve the problem of center point deviation of the K-means + + clustering analysis algorithm, outliers are removed, clustering analysis is performed on data of non-outliers, and a working condition result of each dimension of data, such as working condition 11, working condition 12, working condition 21, and the like in the figure, can be obtained, and a plurality of one-dimensional working condition results are combined into a working condition interval corresponding to each working condition parameter, for example, working condition 11, working condition 12, …, and working condition 1n are combined into working condition 1. By respectively carrying out cluster analysis on each working condition parameter and finding out the working condition interval division of the parameter from the one-dimensional data layer, the problem that the clustering effect of a K-means cluster analysis algorithm on strip-shaped and non-spherical clusters is poor is effectively solved, and the method is also suitable for high-dimensional working condition.
The working condition division method comprises the steps of firstly obtaining each working condition parameter in working condition parameter historical data, carrying out one-dimensional clustering analysis on each working condition parameter, combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis on each working condition parameter into a working condition interval, and determining division of real-time working condition parameters corresponding to working conditions according to the working condition interval.
In the above embodiment, a plurality of clustering analysis results can be obtained by performing one-dimensional K-means + + clustering analysis on each operating condition parameter in the operating condition parameter historical data, and the plurality of clustering analysis results are combined into the operating condition interval, so that the operating condition division method provided by the application can effectively avoid the problem that the clustering effect of the non-spherical clusters and the high-dimensional feature vectors is not obvious. On the basis of the foregoing embodiment, as a preferred embodiment, fig. 4 is a flowchart of a working condition division method provided in another embodiment of the present application; as shown in fig. 4, after acquiring each operating condition parameter in the operating condition parameter history data, the present embodiment further includes: the data is pre-processed.
S110: and eliminating bad point data according to the quality bits of the working condition parameter historical data.
S111: and eliminating overrun data according to the preset upper limit and lower limit of the working condition parameters.
S112: and judging the stability of the working condition parameters according to the stability criterion, and rejecting unstable working condition parameter values in the working condition parameters.
FIG. 5 is a schematic diagram of a thermal power condition division method according to another embodiment of the present application; as shown in fig. 5, the preprocessing of the data is performed by normalization, wherein the normalization is performed to limit the preprocessed data within a certain range, so as to eliminate the adverse effect caused by the singular sample data. In this embodiment, the data is preprocessed differently from the normalization process shown in fig. 5, but by removing dead points and over-limit points and obtaining unstable working condition parameter values according to the stability criterion, the data is preprocessed. In specific implementation, each working condition parameter is preprocessed, and dead point data is removed according to the quality level of the data; according to the preset upper limit and the preset lower limit of each working condition parameter, data exceeding the upper limit and the lower limit, namely overrun data, are removed, wherein the overrun data belong to abnormal working condition data; and acquiring working condition parameters which are continuous for a certain time, and judging the stability of each working condition parameter according to a stability criterion so as to eliminate unstable working condition parameter values.
The specific operation of judging the stability of each working condition parameter according to the stability criterion comprises the following steps:
judging whether each working condition parameter meets the condition that the difference between the maximum value and the minimum value in the data sequence of each working condition parameter is smaller than a preset value, and the quotient of the process deviation amount and the standard deviation of the data sequence of which the number is 3 times is larger than a specified value, if so, determining that the working condition parameters are stable; if not, the unstable working condition parameter value of the working condition parameters can be determined, and the step of rejecting the unstable working condition parameter value in the working condition parameters is carried out.
The stability criterion mentioned in this example is:
wherein, T is a process deviation value which is a user set value, sigma is a standard deviation of a working condition parameter data sequence, max is a maximum value in the data sequence, and min is a minimum value in the data sequence. In specific implementation, the stability of the working condition parameter can be determined only when two conditions meeting the stability criterion are required, and if the working condition parameter does not meet any one of the stability conditions, the working condition parameter value is determined to be unstable, and the working condition parameter value needs to be removed.
It should be noted that, in this embodiment, the determination of the stability of the operating condition parameter by the stability criterion is only a preferred embodiment, and in the specific implementation, the determination method of the stability of the operating condition parameter is not limited.
In the embodiment, each working condition parameter is preprocessed, and dead point data is removed according to the quality level of the data; according to the preset upper limit and lower limit of the working condition parameters, eliminating data which exceed the range of the upper limit and the lower limit, namely overrun data; and judging the stability of the working condition parameters according to the introduced stability criterion, thereby eliminating unstable working condition parameter values, and ensuring that the working condition parameters participating in subsequent clustering analysis are stable working condition parameters.
In the embodiment, before the one-dimensional clustering analysis is performed on each working condition parameter, the data of each working condition parameter is preprocessed, and dead points, over-limit points and unstable working condition parameter values are removed. On the basis of the foregoing embodiment, as a preferred embodiment, after determining the stability of the operating condition parameters according to the stability criterion and removing unstable operating condition parameter values in the operating condition parameters, the method further includes: and carrying out local outlier factor detection algorithm processing on the preprocessed working condition parameter data. The flow of the local outlier detection algorithm is as follows:
s200: determining the kth reachable distance of each point within the kth distance neighborhood of each point.
In this step, the k-th reachable distance of each point in the k-th distance neighborhood of each point is calculated, and the k-th reachable distance is calculated according to the formula: reach _ dist k (o,p)＝max{d k (p), d (o, p) }. Wherein d (o, p) is the distance from the neighborhood point o to the point p, and the one-dimensional data adopts the Manhattan distance d (o, p) = | o-p |. d k (p) the kth distance of the field point o, p is the circle center and radiates outwards until the kth adjacent point q is covered, then d k (p) = d (q, p), and d (q, p) is the distance from the point q to the point p.
S201: a local kth local achievable density for each point is determined.
After the kth reachable distance of each point in the kth distance neighborhood of each point is calculated, the local kth local reachable density of each point is calculated through a formula
And performing calculation determination.
Wherein the content of the first and second substances,
representing the kth reachable distance, N, of each point in the kth distance neighborhood of each point k (p) is the kth distance neighborhood of point p, the kth distance neighborhood of point p N k (p) pointing at the set of all points within the kth distance of p, including the point at the kth distance, | N k (p)|≥k。
S202: the kth local outlier factor for each point is determined.
Calculating the k local outlier factor by the k local reachable distance and the k local reachable density of each point calculated in the steps S200 and S201, and calculating the k local outlier factor by a formula
Calculating to obtain each pointThe k local outlier factor.
S203: and determining the data points to which the maximum n local outlier factors belong, and outputting an outlier set.
In specific implementation, the data points to which the largest n local outlier factors belong are determined by analyzing the k-th local outlier factor of each point obtained through calculation, and a set of outliers is output, i.e., O = { O1, O2, O3, … On }.
According to the embodiment, local factor detection is carried out on the preprocessed working condition parameters, so that the screening of rare working condition parameter data can be realized, and the central point deviation of a K-means cluster analysis algorithm is avoided.
The foregoing embodiment describes in detail the data preprocessing operation and the local outlier detection for each operating condition parameter, respectively, and on the basis of the foregoing embodiment, as a preferred embodiment, fig. 6 is a flowchart of a K-Means + + flowchart provided in another embodiment of the present application; as shown in fig. 6, the steps S300 to S311 of performing K-means + + cluster analysis on the operating condition parameters include: initializing k central points, calculating the distance between each point and the central point, searching the nearest central point to form a cluster, recalculating the average value of the sample data in each cluster, taking the average value as a new cluster center, and judging whether a new cluster center needs to be generated.
In the embodiment, after outliers are removed, K-means + + clustering analysis is performed on the working condition parameter data, and reasonable interval division of the working condition parameters is performed according to the data clustering characteristics of actual working conditions. In the specific implementation, firstly, K initial central points are determined, instead of artificially given central points, which is the largest difference of a K-means cluster analysis method, after the central points are determined, the distances between the ith element and the central points are respectively calculated, so that the central point K closest to the ith element is found to form a cluster, and because the calculation processes are all one-dimensional data, the Manhattan distance formula dist = | x can be used i -c t Completing calculation, after the calculation is completed, bringing the ith element into a cluster k, and after all the elements are traversed, calculating a new central point of each cluster, namely recalculating the average value of sample data in each cluster, namely, by using a formula:
Wherein, c t Represents the center of the t-th cluster, t is more than or equal to 1 and less than or equal to k, | s t I represents the number of objects in the t-th cluster class, x i Represents the ith object in the tth cluster class, 1 is more than or equal to i is less than or equal to | s t L. And finally, comparing the new clustering center with the old clustering center, judging whether the distance is changed or not, namely judging whether the distance is smaller than a threshold value or not, if the new clustering center is judged to be changed compared with the old clustering center, repeatedly executing the steps S302-S309 until the central point is not changed or the maximum cycle number is reached.
FIG. 7 is a flowchart of the K-Means + + algorithm initializing the center point according to another embodiment of the present application; as shown in fig. 7, determining the initialized cluster center includes the following steps:
s400: a point is randomly selected as the first center point.
S401: the distance between each point and the set of central points is calculated, and if the distance is greater than 1, the smallest value is selected, denoted as D (X).
In this step, the distance between each point and the center point needs to be calculated, which is denoted as D (X), and calculated by the manhattan distance formula for the calculated distance described in the above embodiment. It is understood that if the number of center points is greater than 1, each point will have a plurality of distances, and the smallest distance is selected.
S402: a distance probability value P (X) for each point is calculated. By calculation of a formula of probability values
A probability value P (X) is calculated for each distance.
S403: accumulating the distance probability values of each point, wherein the probability values of each point form an interval ranging from 0 to 1.
Accumulating the probability value of each point according to a wheel disc method to generate a probability interval. If there are 8 points, the probability value P (X) of each distance is calculated as shown in table 1, and the accumulated probability value according to the wheel disc method is also shown in the following table.
TABLE 1
| |
① | ② | ③ | ④ | ⑤ | ⑥ | ⑦ | ⑧ |
| P(X) | 0.23 | 0.35 | 0.14 | 0.25 | 0.01 | 0 | 0.005 | 0.015 |
| SUM | 0.23 | 0.58 | 0.72 | 0.97 | 0.98 | 0.98 | 0.985 | 1 |
S404: randomly generating a number in the range of [0.5,1], and determining which interval the number falls in, wherein the point corresponding to the maximum value (the accumulated value to a certain point) of the interval is the next central point.
In the specific implementation, a value in the range of [0.5,1] is randomly generated, and the corresponding point of the maximum value of the accumulated sum interval where the value is located is taken as the next central point. For example, a value in the range of [0.5,1] is randomly generated to be 0.86, and in combination with the accumulated probability value interval in step S403, as shown in table 1, if the random number 0.86 is greater than 0.72 and less than 0.97 and is in the accumulated probability value interval of [0.72,0.97], a corresponding point (4) with a maximum value of 0.97 in the interval is selected, and the point 4 is used as the next central point.
S405: and judging whether the number of the central points is larger than k, if not, returning to the step S401 to calculate the distance between each point and the central point, namely, repeatedly executing the steps S401 to S405 until the number of the central points is k. If yes, go to step S406 to determine that the initialization center point operation is completed.
In the embodiment, the initialization center point is determined by a K-means + + clustering analysis algorithm, and compared with a K-means clustering analysis method used for dividing the current working conditions, the K-means + + does not need to manually determine the initialization clustering center, the number K of the clustering centers does not need to be given in advance, and the initialization center point is random.
The foregoing embodiment describes in detail the K-means + + cluster analysis of one-dimensional operating condition parameters, and on the basis of the foregoing embodiment, as a preferred embodiment, fig. 8 is a flowchart for determining the number of cluster analysis clusters according to another embodiment of the present application; as shown in fig. 8, steps S500-S503: before the clustering analysis result obtained after one-dimensional K-means + + clustering analysis is performed on each working condition parameter to form a working condition interval, the method further comprises the following steps: and determining the number of cluster analysis clusters according to the contour coefficients.
In this embodiment, the number of cluster analysis clusters may be determined with reference to the profile coefficient, and in specific implementation, according to the profile coefficient, the k value recommended in the cluster analysis algorithm is the k value corresponding to the maximum value of the profile coefficient. According to the preset range of a user, traversing K and circulating n times, executing K-means + + clustering analysis with the cluster number equal to K, thereby generating clustering results and K central points, calculating the contour coefficients of the current K, generating n contour coefficients due to the circulation of n times, sequencing the n generated contour coefficients, and selecting the K value corresponding to the maximum value of the contour coefficients to determine the K value as the recommended K value. The method for calculating the contour coefficient comprises the following steps: for one of the points, the point is marked as an i vector, and calculating the contour coefficient corresponding to the point requires calculating a (i) = average, b (i) = min (the distance from the i vector to other points in all the clusters to which the i vector belongs), then the contour coefficient of the i vector is:
k is in the range of preset initial value and maximum value, and if the maximum value is not preset, the maximum value is determined to be
n is the number of samples.
After determining the number of cluster analysis clusters, the embodiment further includes repeatedly performing the operations of preprocessing the operating condition parameters, detecting the outlier factor, initializing the cluster center, and determining the number of clusters of the cluster analysis in the above embodiments until all the parameters are analyzed, and at this time, the operating condition partition rule of each operating condition parameter is generated.
It should be noted that, the operations of preprocessing the operating condition parameters, detecting the outlier factor, initializing the cluster center, and determining the number of clusters in the cluster analysis in the foregoing embodiment are repeated, and may be executed by periodic triggering or manual triggering, which is not limited to this. According to the embodiment, the missing typical working condition intervals in the historical data can be supplemented by repeatedly executing the operations in the embodiment, accurate interval value reference is provided for determining the optimal target value of the operating parameter of the thermal power plant unit of the power plant, the working condition intervals can be updated regularly by executing off-line cluster analysis regularly, and the problems of the working conditions absent in the historical data and the new working conditions generated in real-time operation are solved.
In the embodiment, the contour coefficient is calculated by giving the number k of clusters in an incremental manner in a round-robin manner, the k value corresponding to the maximum value of the contour coefficient is used as the recommended k value, the method is different from the method that the cluster class k in the current clustering analysis algorithm needs to be manually specified, the optimal number of clusters can be automatically determined, and a certain working condition parameter clustering analysis result corresponding to the recommended k value is stored in a working condition interval division configuration table for optimizing the unit, so that a division rule can be provided for the working condition division of real-time working condition parameters.
In the above embodiment, the working condition division method is described in detail, and the application also provides an embodiment corresponding to the working condition division device. It should be noted that the present application describes the embodiments of the apparatus portion from two perspectives, one from the perspective of the function module and the other from the perspective of the hardware.
Based on the angle of the function module, fig. 9 is a structural diagram of a working condition dividing device according to another embodiment of the present application; as shown in fig. 9, the present application provides a condition dividing apparatus, including:
the acquisition module 10 is used for acquiring each working condition parameter in the working condition parameter historical data;
the combination module 11 is used for combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval;
and the dividing module 12 is used for determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
Since the embodiments of the apparatus portion and the method portion correspond to each other, please refer to the description of the embodiments of the method portion for the embodiments of the apparatus portion, which is not repeated here.
The working condition dividing device provided by the application comprises an acquisition module 10, an combination module 11 and a dividing module 12, wherein the acquisition module 10 acquires each working condition parameter in working condition parameter historical data, one-dimensional cluster analysis is carried out on each working condition parameter, the combination module 11 combines a cluster analysis result obtained after each working condition parameter is subjected to one-dimensional K-means + + cluster analysis into a working condition interval, and the dividing module 12 determines the division of real-time working condition parameters corresponding to working conditions according to the working condition interval.
Fig. 10 is a structural diagram of a working condition division apparatus according to another embodiment of the present application; as shown in fig. 10, the condition division apparatus includes: a memory 20 for storing a computer program;
a processor 21 for implementing the steps of the method for division of operating conditions as mentioned in the above embodiments when executing the computer program.
The processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, and the like. The Processor 21 may be implemented in hardware using at least one of a Digital Signal Processor (DSP), a Field-Programmable Gate Array (FPGA), and a Programmable Logic Array (PLA). The processor 21 may also include a main processor and a coprocessor, where the main processor is a processor for Processing data in an awake state, and is also called a Central Processing Unit (CPU); a coprocessor is a low power processor for processing data in a standby state. In some embodiments, the processor 21 may be integrated with a Graphics Processing Unit (GPU) which is responsible for rendering and drawing the content required to be displayed by the display screen. In some embodiments, the processor 21 may further include an Artificial Intelligence (AI) processor for processing computational operations related to machine learning.
The memory 20 may include one or more computer-readable storage media, which may be non-transitory. Memory 20 may also include high speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In this embodiment, the memory 20 is at least used for storing the following computer program 201, wherein after being loaded and executed by the processor 21, the computer program can implement the relevant steps of the operating condition division method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 20 may also include an operating system 202, data 203, and the like, and the storage manner may be a transient storage manner or a permanent storage manner. Operating system 202 may include, among others, windows, unix, linux, and the like. Data 203 may include, but is not limited to, data related to a method of operating condition partitioning, and the like.
In some embodiments, the condition dividing apparatus may further include a display 22, an input/output interface 23, a communication interface 24, a power supply 25, and a communication bus 26.
Those skilled in the art will appreciate that the configuration shown in FIG. 10 does not constitute a limitation of the regime partitioning apparatus and may include more or fewer components than those shown.
The working condition division equipment provided by the embodiment of the application comprises a memory and a processor, wherein when the processor executes a program stored in the memory, the following method can be realized:
the method comprises the steps of firstly obtaining each working condition parameter in working condition parameter historical data, carrying out one-dimensional clustering analysis on each working condition parameter, combining clustering analysis results obtained after each working condition parameter is subjected to one-dimensional K-means + + clustering analysis into working condition intervals, and determining the division of real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
Finally, the application also provides a corresponding embodiment of the computer readable storage medium. The computer-readable storage medium has stored thereon a computer program which, when being executed by a processor, carries out the steps as set forth in the above-mentioned method embodiments.
It is to be understood that if the method in the above embodiments is implemented in the form of software functional units and sold or used as a stand-alone product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which is stored in a storage medium and executes all or part of the steps of the methods described in the embodiments of the present application, or all or part of the technical solutions. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The method, the device, the equipment and the medium for dividing the working conditions provided by the application are described in detail above. The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description. It should be noted that, for those skilled in the art, without departing from the principle of the present application, the present application can also make several improvements and modifications, and those improvements and modifications also fall into the protection scope of the claims of the present application.
It should also be noted that, in this specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
Claims (10)
1. A method for dividing operating conditions, comprising:
acquiring each working condition parameter in the working condition parameter historical data;
combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval;
and determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
2. The operating condition division method according to claim 1, wherein after acquiring each operating condition parameter in the operating condition parameter history data, further comprising:
removing bad point data according to the quality bits of the working condition parameter historical data;
eliminating out-of-limit data according to preset upper and lower limits of the working condition parameters;
and judging the stability of the working condition parameters according to a stability criterion, and rejecting unstable working condition parameter values in the working condition parameters.
3. The operating condition division method according to claim 2, wherein said determining the stability of the operating condition parameters according to the stability criterion comprises:
judging whether the working condition parameters meet the condition that the difference between the maximum value and the minimum value in the data sequence of the working condition parameters is smaller than a preset value, and the quotient of the process deviation amount and 3 times of the standard deviation of the data sequence is larger than a specified value;
if so, determining that the working condition parameters are stable;
if not, determining that unstable working condition parameter values exist in the working condition parameters, and entering the step of rejecting the unstable working condition parameter values in the working condition parameters.
4. The working condition division method according to claim 2, wherein after the stability of the working condition parameters is judged according to the stability criterion and unstable working condition parameter values in the working condition parameters are removed, the method further comprises the following steps:
determining the kth reachable distance of each point in the kth distance neighborhood of each point;
determining a local kth local reachable density of each point;
determining a kth local outlier factor for each point;
and determining the data points to which the maximum n local outlier factors belong, and outputting an outlier set.
5. The operating condition division method according to claim 4, wherein performing a one-dimensional K-means + + clustering analysis on each of the operating condition parameters comprises:
determining an initialization clustering center;
determining the distance between each point and the central point;
determining the average value of the sample data in each cluster as a new cluster center;
judging whether the distance between the new clustering center and the old clustering center is smaller than a threshold value;
if yes, generating a central point and each cluster set;
if not, returning to the step of calculating the distance between each point and the central point.
6. The operating condition partitioning method according to claim 5, wherein the determining of the initialized cluster center comprises:
randomly selecting a point as a first central point;
determining the distance between each point and the central point;
determining a probability value for each distance;
accumulating the probability value of each point according to a wheel disc method to generate a probability interval;
randomly generating a numerical value within a preset range, and taking a maximum value corresponding point of an interval where the numerical value is located as a next central point;
judging whether the number of the central points is greater than k;
if not, returning to the step of calculating the distance between each point and the central point.
7. The operating condition division method according to claim 5, wherein before combining the clustering analysis results obtained after performing one-dimensional K-means + + clustering analysis on each operating condition parameter into an operating condition interval, the method further comprises:
traversing k;
performing K-means + + clustering analysis with the cluster number equal to K to generate clustering results and K central points;
calculating the contour coefficient of the current k to generate a plurality of contour coefficients;
sequencing the contour coefficients, and determining k corresponding to the maximum value as a recommended k value;
storing the clustering analysis result of the working condition parameters corresponding to the recommended k value in a working condition interval division configuration table for unit optimization so as to provide a division rule for the working condition division of the real-time working condition parameters;
wherein the range of k is a preset initial value and a maximum value, and if the maximum value is not preset, the maximum value is determined to be
n is the number of samples.
8. An operation condition division device, comprising:
the acquisition module is used for acquiring each working condition parameter in the working condition parameter historical data;
the combination module is used for combining clustering analysis results obtained after one-dimensional K-means + + clustering analysis is carried out on each working condition parameter into a working condition interval;
and the division module is used for determining the division of the real-time working condition parameters corresponding to the working conditions according to the working condition intervals.
9. An operating condition dividing apparatus comprising a memory for storing a computer program;
a processor for implementing the steps of the regime partitioning method as claimed in any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method of regime partitioning according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211329069.XA CN115618255A (en) | 2022-10-27 | 2022-10-27 | Working condition division method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211329069.XA CN115618255A (en) | 2022-10-27 | 2022-10-27 | Working condition division method, device, equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115618255A true CN115618255A (en) | 2023-01-17 |
Family
ID=84876583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211329069.XA Pending CN115618255A (en) | 2022-10-27 | 2022-10-27 | Working condition division method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115618255A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116304835A (en) * | 2023-03-31 | 2023-06-23 | 北京博华信智科技股份有限公司 | AI-based dynamic equipment working condition monitoring management method, system, equipment and medium |
-
2022
- 2022-10-27 CN CN202211329069.XA patent/CN115618255A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116304835A (en) * | 2023-03-31 | 2023-06-23 | 北京博华信智科技股份有限公司 | AI-based dynamic equipment working condition monitoring management method, system, equipment and medium |
| CN116304835B (en) * | 2023-03-31 | 2023-08-29 | 北京博华信智科技股份有限公司 | AI-based dynamic equipment working condition monitoring management method, system, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108763420B (en) | Data object classification method, device, terminal and computer-readable storage medium | |
| CN114926699B (en) | Indoor three-dimensional point cloud semantic classification method, device, medium and terminal | |
| CN115618255A (en) | Working condition division method, device, equipment and medium | |
| CN111008662B (en) | Online monitoring data anomaly analysis method for power transmission line | |
| CN114861788A (en) | Load abnormity detection method and system based on DBSCAN clustering | |
| CN112911627A (en) | Wireless network performance detection method, device and storage medium | |
| AU2018253470A1 (en) | Device and method for extracting terrain boundary | |
| CN113139880A (en) | Wind turbine generator actual power curve fitting method, device, equipment and storage medium | |
| CN111159900A (en) | Method and device for predicting wind speed of fan | |
| CN114898118A (en) | Automatic statistical method and system for power transmission line house removal amount based on multi-source point cloud | |
| CN114417095A (en) | Data set partitioning method and device | |
| CN107357714B (en) | Fault analysis method and device based on monitoring platform | |
| CN110191005B (en) | Alarm log processing method and system | |
| CN110609832B (en) | Non-repeated sampling method for streaming data | |
| CN113554079B (en) | Power load abnormal data detection method and system based on secondary detection method | |
| CN115794405A (en) | Dynamic resource allocation method of big data processing framework based on SSA-XGboost algorithm | |
| CN112330164B (en) | Data quality management system and method based on message bus | |
| CN107203916B (en) | User credit model establishing method and device | |
| Feng et al. | A genetic k-means clustering algorithm based on the optimized initial centers | |
| CN113778518A (en) | Data processing method, data processing device, computer equipment and storage medium | |
| CN113836826A (en) | Key parameter determination method and device, electronic device and storage medium | |
| Zhu et al. | Effective clustering analysis based on new designed clustering validity index and revised K-means algorithm for big data | |
| CN111698700B (en) | Method and device for judging working state of cell | |
| CN110826904A (en) | Data processing method and device for fan, processing equipment and readable storage medium | |
| CN110825493A (en) | Virtual machine tuning method and device |
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 |