An intelligent identification method for mixed gases based on cluster analysis

Abstract

The present invention relates to the field of gas concentration testing technology, and specifically to a method for intelligently identifying mixed gases based on cluster analysis, including: obtaining the gas concentration of each gas at each moment; clustering the gas concentration of each gas at all moments except the current moment to obtain a number of initial clusters; obtaining the normal evaluation value of the gas mixture state of each initial cluster according to the distribution of the gas concentration in each initial cluster; increasing the gas concentration of each gas at the current moment to cluster to obtain the cluster where the current concentration is located; obtaining the normality of the current moment according to the difference between the normal evaluation values ​​of the cluster where the current concentration is located and the initial cluster; obtaining the possibility of abnormal changes in the gas mixture state at the current moment according to the changes in a period of time closest to the current moment, and performing intelligent abnormal identification on the mixed gas. The present invention identifies the gas mixture state at the current moment by analyzing the gas mixture state at different moments.

Description

Technical Field

[0001] The present invention relates to the technical field of gas concentration testing, and in particular to an intelligent identification method for mixed gases based on cluster analysis. Background Art

[0002] An underground gas storage is an artificial gas field or reservoir formed by re-injecting commercial natural gas transported by long-distance pipelines into underground space. It is generally built near downstream natural gas user cities. Compared with ground spherical tank gas storage, underground gas storage is safer, but there are also some safety issues. When the gas concentration ratio in the gas storage is not appropriate, it will affect the use of the gas in the gas storage or the safety status of the gas storage.

[0003] The traditional method of judging the status of a gas storage reservoir is mainly to set thresholds for the concentrations of multiple gases in the gas tank and perform corresponding processing when the concentration of the gas exceeds the threshold. However, when multiple gases in the gas tank are mixed, the threshold judgment of the concentration of a single gas will be affected, resulting in errors in the judgment of the status of the gas storage reservoir and the inability to control the mixed gas in the gas storage reservoir in a timely manner. Summary of the invention

[0004] The present invention provides a mixed gas intelligent identification method based on cluster analysis to solve the existing problems.

[0005] The present invention provides a method for intelligently identifying mixed gases based on cluster analysis, which adopts the following technical solutions:

[0006] An embodiment of the present invention provides a method for intelligently identifying mixed gases based on cluster analysis, the method comprising the following steps:

[0007] Get the gas concentration of each gas at each moment;

[0008] Clustering the gas concentration of each gas at all times except the current time to obtain several initial clusters; obtaining a normal evaluation value of the gas mixing state of each initial cluster based on the concentration stability distribution of the gas concentration in each initial cluster;

[0009] The gas concentration of each gas at the current moment is increased for clustering to obtain the cluster where the current concentration is located; according to the difference between the normal evaluation values ​​of the gas mixing state of the cluster where the current concentration is located and the initial cluster, the normality of the gas mixing state in the gas storage at the current moment is obtained;

[0010] According to the change of the normal degree of the gas mixture state in a period of time closest to the current moment, the possibility of abnormal change of the gas mixture state at the current moment is obtained;

[0011] According to the possibility of abnormal changes in the gas mixture state at the current moment, intelligent abnormal identification is performed on the mixed gas in the gas storage.

[0012] Furthermore, the step of obtaining the gas concentration of each gas at each moment includes the following specific steps:

[0013] The average value of the concentration of the same gas at each moment collected by all pump-suction composite gas detectors in different preset collection areas of the underground gas storage is used as the concentration of each gas at each moment in the underground gas storage.

[0014] Furthermore, the step of obtaining a plurality of initial clusters includes the following specific steps:

[0015] A two-dimensional sample space is constructed according to the gas concentration of each gas at each moment, in which the horizontal axis represents different types of gas and the vertical axis represents the gas concentration at each moment. The gas concentrations of all types of gas at all moments except the current moment are mapped to the clustering space to form a number of discrete data points. The DBSCAN density clustering algorithm is used to cluster all discrete data points according to the preset clustering radius and the preset number of samples to obtain a number of initial clusters.

[0016] Furthermore, the step of obtaining the normal evaluation value of the gas mixture state of each initial cluster includes the following specific steps:

[0017]

[0018] In the formula, is the normal evaluation value of the gas mixture state of the i-th initial cluster, is the number of data points in the initial cluster, is the dispersion degree of all data points in the i-th initial cluster.

[0019] Furthermore, the specific method of obtaining the dispersion degree is as follows:

[0020]

[0021] In the formula, is the dispersion degree of all data points in the i-th initial cluster, is the Euclidean distance between the two data points with the largest Euclidean distance in the i-th initial cluster, is the area of ​​the region enclosed by the i-th initial cluster in the sample space.

[0022] Furthermore, the step of obtaining the cluster where the current concentration is located includes the following specific steps:

[0023] The different gas concentrations at the current moment are mapped to the existing sample space to form a new sample space, and all data points in the new sample space are re-clustered by DBSCAN to obtain several new initial clusters. The new initial cluster with the current moment data points added to it is recorded as the cluster where the current concentration is located.

[0024] Furthermore, the obtaining of the normality of the gas mixing state in the gas storage at the current moment includes the following specific steps:

[0025] The current moment is recorded as the jth moment, and the calculation method of the normality of the gas mixing state in the gas storage at the jth moment is:

[0026]

[0027] In the formula, is the normality of the gas mixing state in the gas storage at the jth moment, is the number of clusters where the current concentration at the jth moment belongs, is the first The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at time j-1, For the The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment.

[0028] Furthermore, the obtaining of the possibility of abnormal change of the gas mixture state at the current moment includes the following specific steps:

[0029]

[0030] In the formula, is the possibility of abnormal change of the gas mixture state at the jth moment, is the normal increase in the gas mixing state in the gas storage at the kth moment before the jth moment, is the normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment, The number of preset moments.

[0031] Furthermore, the specific method of obtaining the normal degree increase is as follows:

[0032]

[0033] In the formula, is the normal increase in the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the k-1th moment before the jth moment.

[0034] Furthermore, the intelligent abnormal identification of the mixed gas in the gas storage includes the following specific steps:

[0035] The possibility of abnormal change of the gas mixture state at the current moment is normalized to The normalized evaluation of the possibility of abnormal change of the gas mixture state at the current moment is obtained, and an abnormal judgment threshold is preset. When the normalized evaluation of the possibility of abnormal change is greater than the abnormal judgment threshold, a regulator is added to adsorb the gas in the gas storage.

[0036] The beneficial effects of the technical solution of the present invention are: obtaining the gas concentration of each gas at each moment; clustering the gas concentration of each gas at all moments except the current moment to obtain several initial clusters, indicating the gas concentration distribution of the gas in a stable gas mixture state; obtaining the normal evaluation value of the gas mixture state of each initial cluster according to the distribution of the gas concentration in each initial cluster; increasing the gas concentration of each gas at the current moment to cluster and obtain the cluster where the current concentration is located; obtaining the normality of the current moment according to the difference between the normal evaluation values ​​of the cluster where the current concentration is located and the initial cluster, indicating the gas mixture state in the gas storage after the current moment is increased; obtaining the possibility of abnormal changes in the gas mixture state at the current moment according to the changes in a period of time closest to the current moment, and timely performing intelligent abnormal identification on the mixed gas. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 The present invention is a flow chart of the steps of a method for intelligently identifying mixed gases based on cluster analysis. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, characteristics and effects of a mixed gas intelligent identification method based on cluster analysis proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0041] The specific scheme of the mixed gas intelligent identification method based on cluster analysis provided by the present invention is described in detail below with reference to the accompanying drawings.

[0042] See also Figure 1 , which shows a flow chart of a method for intelligently identifying mixed gases based on cluster analysis provided by an embodiment of the present invention, the method comprising the following steps:

[0043] Step S001: Obtain the gas concentration of each gas at each moment.

[0044] The purpose of this embodiment is to analyze the gas mixture state at different times to obtain the gas mixture state at the current moment; therefore, firstly, multiple pump-suction composite gas detectors are installed in the underground gas storage to detect the concentrations of methane, ethane, propane, and butane gases in the main gases at each moment; the average value of the same gas concentration at each moment collected by all pump-suction composite gas detectors in different preset collection areas of the underground gas storage is used as the concentration of each gas at each moment in the underground gas storage, and the unit of the gas concentration is ; In this embodiment, a pump-suction composite gas detector is installed every ten meters, and the concentration of different gases in the gas storage reservoir is detected every 10 minutes. Other installation distances and detection frequencies can be set in other implementation methods, and this embodiment does not specifically limit them.

[0045] Step S002: cluster the gas concentration of each gas at all times except the current time to obtain a number of initial clusters; obtain a normal evaluation value of the gas mixing state of each initial cluster according to the concentration stability distribution of the gas concentration in each initial cluster.

[0046] It should be noted that when the concentration of flammable gases such as methane in underground gas storage is high, explosions and other conditions are more likely to occur in the underground gas storage; when the concentration of propane and ethane is high, liquid oil is easily generated and deposited on pipelines and equipment, causing blockages and other safety hazards. Therefore, it is necessary to analyze different gas mixture states, and different gas mixture states can be clustered using the DBSCAN algorithm.

[0047] A two-dimensional sample space is constructed according to the gas concentration of each gas at each moment, wherein the horizontal axis represents different types of gas and the vertical axis represents the gas concentration at each moment. The gas concentrations of all types of gas at all moments except the current moment are mapped to the clustering space to form a number of discrete data points. The DBSCAN density clustering algorithm is used to cluster all discrete data points according to the preset clustering radius and the preset number of samples to obtain a number of initial clusters. Each initial cluster represents the concentration distribution of a gas in the same stable mixed state. It should be noted that in this embodiment, the preset clustering radius is 7. , the clustering sample number is 20 for example. The DBSCAN density clustering algorithm is a well-known technology. Other embodiments may use other methods, which will not be described in detail in this embodiment.

[0048] It should be further explained that after clustering all the data points, each initial cluster represents a gas mixture state, indicating the gas concentration distribution of the gas in a stable gas mixture state; under normal conditions of the gas mixture state in the gas storage, the gas mixture state in the gas storage is normal for a longer period of time, so the more data points in the initial cluster corresponding to the normal state, the denser; but when the gas mixture state in the gas storage is abnormal, the time for maintaining a stable gas concentration is relatively short, and the concentration of each gas changes relatively quickly during the abnormal state, so the data points in the cluster corresponding to the abnormal state are fewer, more dispersed, and the enclosed area is larger.

[0049] Specifically, the degree of dispersion of all data points in the i-th initial cluster is calculated as follows:

[0050]

[0051] The calculation method of the normal evaluation value of the gas mixture state of the i-th initial cluster is:

[0052]

[0053] In the formula, is the normal evaluation value of the gas mixture state of the i-th initial cluster, is the Euclidean distance between the two data points with the largest Euclidean distance in the i-th initial cluster, is the number of data points in the initial cluster, is the area of ​​the region enclosed by the i-th initial cluster in the sample space, is the dispersion degree of all data points in the i-th initial cluster.

[0054] when The smaller it is, the closer the distance between the data points in the i-th initial cluster is, the denser the distribution is, and the more normal the gas mixing state is; The larger the number of initial clusters, the larger the area of ​​the region enclosed by the i-th initial cluster in the sample space. The smaller it is, the more concentrated the data points in the initial cluster are and the more normal the gas mixing state is.

[0055] Step S003: adding the gas concentration of each gas at the current moment to cluster, and obtaining the normality of the gas mixing state in the gas storage at the current moment according to the difference in the normal evaluation value of the gas mixing state of the cluster at the current moment and the cluster at the historical moment.

[0056] It should be noted that when identifying the gas mixture state at the current moment, the gas mixture state at the current moment can be analyzed by comparing it with the historical gas mixture state. If the gas mixture state at the current moment is more consistent with the gas mixture state of the historical normal state, it means that the gas mixture state of the gas storage at the current moment is more normal.

[0057] Specifically, different gas concentrations at the current moment are mapped to the existing sample space to form a new sample space, and DBSCAN clustering is performed again on all data points in the new sample space to obtain several new initial clusters. The new initial cluster with the current moment data points added to it is recorded as the cluster where the current concentration is located.

[0058] It should be further explained that if the data points at the current moment are clustered as a cluster with a smaller normal evaluation value after clustering, it means that the gas mixing state in the gas storage at the current moment is more abnormal; if the data points at the current moment are clustered as a cluster with a higher normal evaluation value after clustering, it means that the gas mixing state in the gas storage at the current moment is more normal.

[0059] Furthermore, the current moment is recorded as the jth moment, and the calculation method of the normality of the gas mixing state in the gas storage at the jth moment is:

[0060]

[0061] In the formula, is the normality of the gas mixing state in the gas storage at the jth moment, is the number of clusters where the current concentration at the jth moment belongs, is the first The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at time j-1, For the The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment.

[0062] In order to increase the change difference of the normal evaluation value in the cluster corresponding to the data point at the current moment, the larger the difference value is, the more normal the gas mixing state in the gas storage at the current moment is.

[0063] Step S004: Obtain the possibility of abnormal change of the gas mixture state at the current moment according to the change of the normal degree of the gas mixture state between adjacent moments in a period of time before the current moment.

[0064] It should be noted that the change in the concentration of mixed gas in the gas storage requires a certain process and time. Therefore, when judging whether the current gas mixture state is normal, it is necessary to consider the change in the gas mixture state in the period before the current moment. If the normal degree of the gas mixture state at the current moment is reduced compared with the normal degree of the mixed gas state at the previous moments, it means that the gas mixture state in the gas storage has become more and more abnormal during this period of time, and the smaller the normal evaluation value of the gas mixture state of the cluster where the current concentration is located is compared with other clusters, the more abnormal the gas mixture state at the current moment is, and the more likely it is that there will be dangers such as explosion, and the gas in the gas storage needs to be processed.

[0065] Specifically, the normal increase in the gas mixing state in the gas storage at the jth moment is:

[0066]

[0067] The calculation method of the abnormal change possibility of the gas mixture state at the jth moment is:

[0068]

[0069] In the formula, is the possibility of abnormal change of the gas mixture state at the jth moment, is the normal increase in the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the k-1th moment before the jth moment, is the normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment, It is the number of preset moments. In this embodiment, the number of preset moments is 8. This is taken as an example for description. Other values ​​may be set in other embodiments, and this embodiment does not specifically limit it.

[0070] is the change in the normal degree of the gas mixture state in the gas storage at two moments. When the difference is larger, the possibility that the gas mixture state in the gas storage at two moments changes to an abnormal state is greater, and when The smaller the value is, the smaller the normal evaluation value of the gas mixture state of the cluster at the jth moment is, and the more likely the gas mixture state at the current moment is abnormal, and the gas in the gas storage needs to be processed accordingly.

[0071] At this point, the possibility of abnormal change in the gas mixture state at the current moment is obtained.

[0072] Step S005: Identify the mixed gas in the gas storage according to the possibility of abnormal change of the gas mixture state at the current moment.

[0073] The possibility of abnormal change of the gas mixture state at the current moment is normalized to The normalized evaluation of the possibility of abnormal change of the gas mixture state at the current moment is obtained, and an abnormal judgment threshold is preset. When the normalized evaluation of the possibility of abnormal change is greater than the abnormal judgment threshold, it indicates that the gas in the gas storage at the current moment may be abnormal, and it is necessary to add a regulator to adsorb the gas with a high concentration in the gas storage and process the mixed gas in the gas storage. In this embodiment, the preset abnormal judgment threshold is 0.5, which is used as an example for description. Other values ​​can be set in other embodiments, and this embodiment does not specifically limit it.

[0074] So far, the present invention is completed.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for intelligent identification of mixed gases based on cluster analysis, characterized in that: The method comprises the following steps: Get the gas concentration of each gas at each moment; Clustering the gas concentration of each gas at all times except the current time to obtain several initial clusters; obtaining a normal evaluation value of the gas mixing state of each initial cluster based on the concentration stability distribution of the gas concentration in each initial cluster; The specific steps of obtaining a plurality of initial clusters are as follows: A two-dimensional sample space is constructed according to the gas concentration of each gas at each moment, wherein the horizontal axis represents different types of gas and the vertical axis represents the gas concentration at each moment. The gas concentrations of all types of gas at all moments except the current moment are mapped to the clustering space to form a number of discrete data points. The DBSCAN density clustering algorithm is used to cluster all discrete data points according to the preset clustering radius and the preset number of samples to obtain a number of initial clusters. The specific steps of obtaining the normal evaluation value of the gas mixing state of each initial cluster are as follows: In the formula, is the normal evaluation value of the gas mixture state of the i-th initial cluster, is the number of data points in the initial cluster, is the dispersion degree of all data points in the i-th initial cluster; The specific method of obtaining the dispersion degree is as follows: In the formula, is the dispersion degree of all data points in the i-th initial cluster, is the Euclidean distance between the two data points with the largest Euclidean distance in the i-th initial cluster, is the area of ​​the region enclosed by the i-th initial cluster in the sample space; The gas concentration of each gas at the current moment is increased for clustering to obtain the cluster where the current concentration is located; according to the difference between the normal evaluation values ​​of the gas mixing state of the cluster where the current concentration is located and the initial cluster, the normality of the gas mixing state in the gas storage at the current moment is obtained; According to the change of the normal degree of the gas mixture state in a period of time closest to the current moment, the possibility of abnormal change of the gas mixture state at the current moment is obtained; The specific steps of obtaining the possibility of abnormal change of the gas mixture state at the current moment are as follows: In the formula, is the possibility of abnormal change of the gas mixture state at the jth moment, is the normal increase in the gas mixing state in the gas storage at the kth moment before the jth moment, is the normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment, is the preset number of moments; According to the possibility of abnormal changes in the gas mixture state at the current moment, intelligent abnormal identification is performed on the mixed gas in the gas storage.

2. The method for intelligently identifying mixed gases based on cluster analysis according to claim 1, characterized in that: The specific steps of obtaining the gas concentration of each gas at each moment are as follows: The average value of the concentration of the same gas at each moment collected by all pump-suction composite gas detectors in different preset collection areas of the underground gas storage is used as the concentration of each gas at each moment in the underground gas storage.

3. The method for intelligently identifying mixed gases based on cluster analysis according to claim 1, characterized in that: The specific steps of obtaining the cluster where the current concentration is located are as follows: The different gas concentrations at the current moment are mapped to the existing sample space to form a new sample space, and all data points in the new sample space are re-clustered by DBSCAN to obtain several new initial clusters. The new initial cluster with the current moment data points added to it is recorded as the cluster where the current concentration is located.

4. The method for intelligently identifying mixed gases based on cluster analysis according to claim 1, characterized in that: The specific steps of obtaining the normality of the gas mixing state in the gas storage at the current moment are as follows: The current moment is recorded as the jth moment, and the calculation method of the normality of the gas mixing state in the gas storage at the jth moment is: In the formula, is the normality of the gas mixing state in the gas storage at the jth moment, is the number of clusters where the current concentration at the jth moment belongs, is the first The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at time j-1, For the The normal evaluation value of the gas mixture state of the cluster where the current concentration is located at the jth moment.

5. The method for intelligently identifying mixed gases based on cluster analysis according to claim 1, characterized in that: The specific method of obtaining the normal degree increase is as follows: In the formula, is the normal increase in the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the kth moment before the jth moment, is the normality of the gas mixing state in the gas storage at the k-1th moment before the jth moment.

6. The method for intelligently identifying mixed gases based on cluster analysis according to claim 1, characterized in that: The intelligent abnormal identification of the mixed gas in the gas storage includes the following specific steps: The possibility of abnormal change of the gas mixture state at the current moment is normalized to The normalized evaluation of the possibility of abnormal change of the gas mixture state at the current moment is obtained, and an abnormal judgment threshold is preset. When the normalized evaluation of the possibility of abnormal change is greater than the abnormal judgment threshold, a regulator is added to adsorb the gas in the gas storage.

Images