AI-based precise prediction algorithm for aeration quantity

Abstract

The application relates to an AI-based precise prediction algorithm for an aeration amount, and belongs to the field of intelligent water affairs, comprising the following steps: acquiring historical data of multiple operation indexes of a biochemical pool, calculating cosine similarity between time sequence vectors of the multiple operation indexes, grouping time sequence vectors with cosine similarity not exceeding a first threshold value to obtain a plurality of first vector groups; acquiring time sequence vectors adjacent to the time sequence vectors of the first time sequence vector group to obtain a second vector group; acquiring a time sequence vector composed of multiple operation indexes of adjacent time periods on the left side of a time axis of a to-be-measured time period to obtain a third vector group; determining corresponding first vector groups of the third vector group according to the cosine similarity between each time sequence vector in the third vector group and the time sequence vectors in the plurality of first vector groups; and determining the aeration amount of the to-be-measured time period according to the corresponding first vector groups of the third vector group. The aeration amount output by the algorithm has good accuracy, can help reduce aeration energy consumption, and promotes low-carbon development of sewage treatment.

Description

Technical Field

[0001] This application relates to the field of smart water technology, and in particular to an AI-based algorithm for accurate prediction of aeration volume. Background Technology

[0002] Aeration refers to the use of equipment such as blowers and aeration pipes to introduce air into a liquid to increase the dissolved oxygen (DO) in the liquid. In the process of water purification in water treatment plants, the main functions of aeration include: (1) increasing the oxidation rate: aerobic microorganisms consume oxygen to degrade organic matter. Aeration increases the dissolved oxygen content in the water, thereby increasing the oxidation rate. In the presence of aerobic microorganisms, aeration can promote the decomposition and metabolism of matter. By changing the dissolved oxygen (DO) in the water, ammonia nitrogen (NH4) in the wastewater is converted into nitrate nitrogen (NO3), while degrading organic matter in the water; (2) promoting microbial mixing: aeration mixes the matter and other suspended solids in the wastewater thoroughly through stirring, making them evenly distributed. This can increase the contact opportunities between microorganisms and organic matter, thereby accelerating the degradation process; (3) sedimentation: aeration stirs the water body, which can cause suspended solids and microorganisms to settle in the water, thereby maintaining the clarity of the water. Although aeration has the above-mentioned effects, the energy consumption of the aeration process accounts for 30% to 70% of the total energy consumption of the wastewater treatment plant. It is energy-intensive and not conducive to economic and environmental protection. Therefore, it is necessary to control the aeration rate while ensuring the quality of the effluent.

[0003] Traditional methods for controlling aeration volume rely heavily on experience and trial and error. However, in actual production, aeration volume data fluctuates significantly and lacks strong periodicity. Furthermore, predicting aeration volume involves numerous influencing factors, making accurate prediction difficult. This method is not only inefficient but also struggles to balance treatment effectiveness and energy consumption. Currently, precise control of aeration volume is not feasible in actual production. Summary of the Invention

[0004] To address the aforementioned issues, this application provides an AI-based algorithm for accurate prediction of aeration volume, comprising:

[0005] Historical data of multiple operating indicators of the biological treatment tank are obtained, including aeration rate.

[0006] The historical data of the aforementioned multiple operational indicators are preprocessed to obtain the time-series vectors of the aforementioned multiple operational indicators;

[0007] Calculate the cosine similarity between the time-series vectors of the multiple operating indicators, and group the time-series vectors whose cosine similarity does not exceed a first threshold to obtain multiple first time-series vector groups;

[0008] Obtain a time-series vector that is located to the right of the first time-series vector group and whose time distance from at least one time-series vector in the time axis does not exceed a time distance threshold; thus, obtain a second time-series vector group corresponding to the first time-series vector group.

[0009] Obtain the time series vector composed of the multiple operating indicators of the time period adjacent to the left of the time axis during the period to be tested, and obtain the third time series vector group;

[0010] Calculate the cosine similarity between each time vector in the third time vector group and the time vectors in the plurality of first time vector groups, and determine at least one first time vector group corresponding to the third time vector group;

[0011] At least one first time-series vector group corresponding to the third time-series vector group is determined as the fourth time-series vector group;

[0012] The aeration rate for the time period to be measured is determined based on the second time series vector group corresponding to the fourth time series vector group.

[0013] Specifically, the aforementioned multiple operating indicators also include influent organic matter concentration, influent ammonia nitrogen concentration, aerobic activated sludge concentration, and dissolved oxygen concentration.

[0014] Specifically, the historical data of the multiple operational indicators are preprocessed to obtain the time-series vectors of the multiple operational indicators, including the following steps:

[0015] Based on the time of historical data, the historical data of the multiple operational indicators are calibrated on a time axis; abnormal and erroneous data in the historical data of the multiple operational indicators are removed; the missing data of the multiple operational indicators is filled by interpolation based on the adjacent data of the same type. The preprocessed historical data of the multiple operational indicators are then arranged into multiple time-series vectors according to their positions on the time axis.

[0016] Specifically, time-series vectors with cosine similarity not exceeding a first threshold are grouped together to obtain multiple first time-series vector groups, including:

[0017] Arrange the time-series vectors of the multiple operating indicators in descending order of cosine similarity. Group two time-series vectors whose cosine similarity does not exceed the first threshold into the same group. Group the time-series vectors whose cosine similarity exceeds the first threshold into a separate group to obtain multiple first time-series vector groups.

[0018] Specifically, calculating the cosine similarity between each time-series vector in the third time-series vector group and the time-series vectors in the plurality of first time-series vector groups, and determining at least one first time-series vector group corresponding to the third time-series vector group, includes:

[0019] Calculate the cosine similarity between each time vector in the third time vector group and each time vector in the plurality of first time vector groups to obtain the cosine similarity matrix of each time vector in the third time vector group.

[0020]

[0021] Represents the third time-series vector group m The time series vector and the first k The matrix formed by the cosine similarity of each time vector in the first time vector group. m This indicates the number of time vectors in the third time vector group. n Indicates the first k The number of time vectors in the first time vector group Represents the third time-series vector group's first... i The time series vector and the first k The first time-series vector group j Cosine similarity of time-series vectors i This indicates the time vector index of the third time vector group. j Indicates the first k The time sequence vector index of the first time sequence vector group;

[0022] Summing each term in the cosine similarity matrix yields the th term. k The matching degree of the first time series vector group to the third time series vector group. , Indicates the first k The matching degree of the first time-series vector group to the third time-series vector group; the first time-series vector group whose matching degree exceeds the second threshold is determined as the first time-series vector group corresponding to the third time-series vector group.

[0023] Specifically, if the number of first time-series vector groups corresponding to the third time-series vector group is not less than 3, then at least one first time-series vector group corresponding to the third time-series vector group is determined as the fourth time-series vector group, including:

[0024] Calculate the first center time-series vector of each first time-series vector group corresponding to the third time-series vector group;

[0025] Calculate the cosine similarity between the first central time-series vectors of each first time-series vector group corresponding to the third time-series vector group;

[0026] The first time-series vector group corresponding to the two first center time-series vectors with the highest cosine similarity is determined as the fourth time-series vector group.

[0027] Specifically, if the number of second time-series vector groups corresponding to the fourth time-series vector group is not less than 3, then the aeration rate for the time period to be measured is determined based on the second time-series vector groups corresponding to the fourth time-series vector group, including:

[0028] Calculate the second center time-series vector of each second time-series vector group corresponding to the fourth time-series vector group;

[0029] Calculate the cosine similarity between the second central time-series vectors of each second time-series vector group corresponding to the fourth time-series vector group;

[0030] The aeration rate for the time period to be measured is determined based on the second time series vector group corresponding to the two second center time series vectors with the highest cosine similarity.

[0031] Specifically, the aeration rate for the time period to be measured is determined based on the second time-series vector group corresponding to the two second center time-series vectors with the highest cosine similarity, including:

[0032] The aeration amount of the time series vector that is closest in time distance to the time period to be measured is determined from the two time series vectors with the highest cosine similarity corresponding to the second time series vectors of the second center.

[0033] The beneficial effects of the above-mentioned technical solutions provided in this application include at least the following:

[0034] The algorithm provided in this application combines the process flow and historical data of the wastewater treatment plant with big data to output the aeration volume for the test period. This helps to control the aeration blower to avoid providing more air volume than predicted during the test period, thus avoiding excessive energy consumption. By adopting intelligent and automated methods to reduce energy consumption, it helps to improve the accuracy of aeration volume prediction and contributes to the low-carbon development of the municipal wastewater treatment industry.

[0035] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0036] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0037] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. In the drawings:

[0038] Figure 1 This is a flowchart of the AI-based accurate prediction algorithm for aeration volume in an embodiment of this application;

[0039] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0040] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0041] To address the problems existing in the prior art, this application provides an AI-based algorithm for accurate prediction of aeration volume, the process of which is as follows: Figure 1 As shown, it includes the following steps:

[0042] Step 1: Obtain historical data of multiple operating indicators of the biochemical tank, including aeration rate; preprocess the historical data of the multiple operating indicators to obtain the time-series vectors of the multiple operating indicators.

[0043] In some specific embodiments, the multiple operating indicators also include influent organic matter concentration, influent ammonia nitrogen concentration, aerobic activated sludge concentration, and dissolved oxygen concentration.

[0044] In some specific embodiments, the historical data of the multiple operational indicators are preprocessed to obtain the time-series vectors of the multiple operational indicators, including:

[0045] Based on the time of historical data, the historical data of the multiple operational indicators are calibrated on a time axis; abnormal and erroneous data in the historical data of the multiple operational indicators are removed; the missing data of the multiple operational indicators is filled by interpolation based on the adjacent data of the same type. The preprocessed historical data of the multiple operational indicators are then arranged into multiple time-series vectors according to their positions on the time axis.

[0046] Step 2: Calculate the cosine similarity between the time-series vectors of the multiple operational indicators, group the time-series vectors whose cosine similarity does not exceed a first threshold to obtain multiple first time-series vector groups; obtain a time-series vector on the time axis that is not more than a time distance threshold from at least one time-series vector of the first time-series vector group and is located to the right of the time-series vector, to obtain a second time-series vector group corresponding to the first time-series vector group; obtain the time-series vector composed of the multiple operational indicators of the time period adjacent to the left of the time axis during the test period to obtain a third time-series vector group.

[0047] In some specific embodiments, time-series vectors with cosine similarity not exceeding a first threshold are grouped together to obtain multiple first time-series vector groups, including:

[0048] Arrange the time-series vectors of the multiple operating indicators in descending order of cosine similarity. Group two time-series vectors whose cosine similarity does not exceed the first threshold into the same group. Group the time-series vectors whose cosine similarity exceeds the first threshold into a separate group to obtain multiple first time-series vector groups.

[0049] Further, obtaining a second time-series vector group corresponding to the first time-series vector group by acquiring a time-series vector that is no more than a time distance threshold from at least one time-series vector in the first time-series vector group and is located to the right of the time-series vectors, including:

[0050] Within the immediate time range (within the time distance threshold) of each time vector in the first time vector group and after the time of that time vector, select multiple time vectors to obtain the second time vector group corresponding to the first time vector group.

[0051] Furthermore, the time series vector composed of the multiple operating indicators of the time period adjacent to the left of the time axis of the time period to be tested is obtained to obtain the third time series vector group, including: selecting multiple time series adjacent to the time period before the time period to be tested to obtain the third time series vector group.

[0052] Step 3: Calculate the cosine similarity between each time vector in the third time vector group and the time vectors in the plurality of first time vector groups, and determine at least one first time vector group corresponding to the third time vector group; determine at least one first time vector group corresponding to the third time vector group as the fourth time vector group; determine the aeration rate for the period to be measured based on the second time vector group corresponding to the fourth time vector group.

[0053] In some specific embodiments, calculating the cosine similarity between each time vector in the third time vector group and the time vectors in the plurality of first time vector groups, and determining at least one first time vector group corresponding to the third time vector group, includes:

[0054] Calculate the cosine similarity between each time vector in the third time vector group and each time vector in the plurality of first time vector groups to obtain the cosine similarity matrix of each time vector in the third time vector group.

[0055]

[0056] Represents the third time-series vector group m The time series vector and the first kThe matrix formed by the cosine similarity of each time vector in the first time vector group. m This indicates the number of time vectors in the third time vector group. n Indicates the first k The number of time vectors in the first time vector group Represents the third time-series vector group's first... i The time series vector and the first k The first time-series vector group j Cosine similarity of time-series vectors i This indicates the time vector index of the third time vector group. j Indicates the first k The time sequence vector index of the first time sequence vector group;

[0057] Summing each term in the cosine similarity matrix yields the th term. k The matching degree of the first time series vector group to the third time series vector group. , Indicates the first k The matching degree of the first time-series vector group to the third time-series vector group; the first time-series vector group whose matching degree exceeds the second threshold is determined as the first time-series vector group corresponding to the third time-series vector group.

[0058] In some specific embodiments, if the number of first time-series vector groups corresponding to the third time-series vector group is not less than 3, then at least one first time-series vector group corresponding to the third time-series vector group is determined as the fourth time-series vector group, including:

[0059] Calculate the first center time-series vector of each first time-series vector group corresponding to the third time-series vector group;

[0060] Calculate the cosine similarity between the first central time-series vectors of each first time-series vector group corresponding to the third time-series vector group;

[0061] The first time-series vector group corresponding to the two first center time-series vectors with the highest cosine similarity is determined as the fourth time-series vector group.

[0062] If the number of first time series vector groups corresponding to the third time series vector group is 1 or 2, then the first time series vector group corresponding to the third time series vector group is determined as the fourth time series vector group.

[0063] Furthermore, by calculating the mean of each data item in the first time series vector group, the first central time series vector of the first time series vector group is determined. This method is used to determine the first central time series vector of each first time series vector group corresponding to the third time series vector group.

[0064] In some specific embodiments, if the number of second time-series vector groups corresponding to the fourth time-series vector group is not less than 3, then the aeration rate for the time period to be measured is determined based on the second time-series vector groups corresponding to the fourth time-series vector group, including:

[0065] Calculate the second center time-series vector of each second time-series vector group corresponding to the fourth time-series vector group;

[0066] Calculate the cosine similarity between the second central time-series vectors of each second time-series vector group corresponding to the fourth time-series vector group;

[0067] The aeration rate for the time period to be measured is determined based on the second time series vector group corresponding to the two second center time series vectors with the highest cosine similarity.

[0068] If the number of second time series vector groups corresponding to the fourth time series vector group is 1 or 2, then the aeration amount for the period to be measured is determined according to the second time series vector group corresponding to the fourth time series vector group.

[0069] Furthermore, by calculating the mean of each data item of the second time series vector group, the second center time series vector of the second time series vector group is determined. The second center time series vector of each second time series vector group corresponding to the fourth time series vector group is determined by this method.

[0070] By using the above methods, we can continuously approach historical data that are the same or similar to the conditions of the period to be tested, and make more accurate predictions about the aeration volume of the period to be tested based on historical data under the same conditions.

[0071] In some specific embodiments, the aeration rate for the time period to be measured is determined based on the second time series vector group corresponding to the two second center time series vectors with the highest cosine similarity, including:

[0072] The aeration rate of the time series vector closest in time distance to the time period to be measured is determined from the two time series vectors corresponding to the two second center time series vectors with the highest cosine similarity. By continuously approaching historical data that are the same or similar to the time period to be measured, it is helpful to make more accurate predictions of the aeration rate of the time period to be measured.

[0073] In the above method of this embodiment, the process flow and historical data of the sewage treatment plant are combined, and big data is used to approximate the historical data of the same or similar situation in the test period through the cosine similarity between time series vectors. The aeration volume of the test period is output, which helps to control the aeration blower to avoid providing air volume beyond the predicted aeration volume in the test period, thus avoiding excessive energy consumption. The use of intelligent and automated means to reduce energy consumption helps to improve the accuracy of aeration volume prediction and helps to achieve low-carbon development in the municipal sewage treatment industry.

[0074] Those skilled in the art can change the above order without departing from the scope of protection of this application.

[0075] Another embodiment of this application provides an AI-based system for accurate prediction of aeration volume, including:

[0076] The historical data preprocessing module is used to acquire historical data of multiple operating indicators of the biochemical tank, including aeration rate; and to preprocess the historical data of the multiple operating indicators to obtain the time-series vectors of the multiple operating indicators.

[0077] The intermediate data processing module is used to calculate the cosine similarity between the time-series vectors of the multiple operational indicators, group the time-series vectors whose cosine similarity does not exceed a first threshold to obtain multiple first time-series vector groups; obtain a time-series vector that is located to the right of at least one time-series vector in the first time-series vector group on the time axis and does not exceed a time distance threshold, to obtain a second time-series vector group corresponding to the first time-series vector group; and obtain the time-series vector composed of the multiple operational indicators in the time-series adjacent to the time-series on the left side of the time axis for the test period to obtain a third time-series vector group.

[0078] The aeration volume prediction module is used to calculate the cosine similarity between each time-series vector in the third time-series vector group and the time-series vectors in the plurality of first time-series vector groups, to determine at least one first time-series vector group corresponding to the third time-series vector group; to determine the at least one first time-series vector group corresponding to the third time-series vector group as the fourth time-series vector group; and to determine the aeration volume for the period to be measured based on the second time-series vector group corresponding to the fourth time-series vector group.

[0079] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0080] In this embodiment, by combining the wastewater treatment plant process flow and historical data, and utilizing big data, the aeration volume for the test period is output through an algorithm. This helps to control the aeration blower to avoid providing more air volume than predicted during the test period, thus avoiding excessive energy consumption. By adopting intelligent and automated methods to reduce energy consumption, the accuracy of aeration volume prediction is improved, which helps to achieve low-carbon development in the municipal wastewater treatment industry.

[0081] Based on the same inventive concept, embodiments of this application also provide an electronic device, the structure of which is as follows: Figure 2 As shown, it includes: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the aforementioned steps of the AI-based aeration volume accurate prediction algorithm.

[0082] Based on the same inventive concept, this application also provides a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the aforementioned steps of the AI-based aeration volume accurate prediction algorithm.

[0083] Based on the same inventive concept, this application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the aforementioned steps of the AI-based aeration volume accurate prediction algorithm.

[0084] Any modifications, additions, and equivalent substitutions made within the scope of the principles of this application shall still fall within the scope of the patent coverage of this application.

[0085] The terms "first" and "second" mentioned above do not indicate a sequential order, but rather represent a distinction between different characteristics.

Claims

1. An AI-based precise prediction algorithm for aeration quantity, characterized in that, The method comprises: obtaining historical data of multiple operation indexes of a biochemical pool, the multiple operation indexes comprising an aeration amount; preprocessing the historical data of the multiple operation indexes to obtain time sequence vectors of the multiple operation indexes; calculating cosine similarities between the time sequence vectors of the multiple operation indexes, and grouping time sequence vectors with cosine similarities not exceeding a first threshold to obtain multiple first time sequence vector groups; obtaining time sequence vectors on a time axis that are not more than a time distance threshold from at least one time sequence vector of the first time sequence vector groups and located on the right side of the time sequence vector to obtain second time sequence vector groups corresponding to the first time sequence vector groups; obtaining time sequence vectors of the multiple operation indexes of a time period adjacent to a left side of a to-be-tested time period on a time axis to obtain third time sequence vector groups; calculating cosine similarities between each time sequence vector in the third time sequence vector groups and time sequence vectors in the multiple first time sequence vector groups to determine at least one first time sequence vector group corresponding to the third time sequence vector groups; determining the at least one first time sequence vector group corresponding to the third time sequence vector groups as fourth time sequence vector groups; determining the aeration amount of the to-be-tested time period according to the second time sequence vector groups corresponding to the fourth time sequence vector groups.

2. The AI-based precise prediction algorithm for aeration quantity according to claim 1, wherein, The multiple operation indexes further comprise influent organic matter concentration, influent ammonia nitrogen concentration, aerobic active sludge concentration, and dissolved oxygen concentration. 3.The AI-based precise prediction algorithm for aeration quantity according to claim 1, wherein, The preprocessing of the historical data of the multiple operation indexes to obtain time sequence vectors of the multiple operation indexes comprises: time axis calibration of the historical data of the multiple operation indexes according to time of the historical data; elimination of abnormal data and error data in the historical data of the multiple operation indexes; interpolation filling of missing data of the multiple operation indexes according to adjacent same type data of the missing data; and composition of multiple time sequence vectors from the preprocessed historical data of the multiple operation indexes according to positions on a time axis.

4. The AI-based precise prediction algorithm for aeration quantity according to claim 1, wherein, The grouping of time sequence vectors with cosine similarities not exceeding a first threshold to obtain multiple first time sequence vector groups comprises: arranging cosine similarities between time sequence vectors of the multiple operation indexes from large to small, grouping two time sequence vectors with cosine similarities not exceeding a first threshold into the same group, and grouping time sequence vectors with cosine similarities all exceeding a first threshold into a group to obtain multiple first time sequence vector groups. 5.The AI-based precise prediction algorithm of aeration quantity according to claim 1, wherein, The calculation of cosine similarities between each time sequence vector in the third time sequence vector groups and time sequence vectors in the multiple first time sequence vector groups to determine at least one first time sequence vector group corresponding to the third time sequence vector groups comprises: calculating cosine similarities between each time sequence vector in the third time sequence vector groups and each time sequence vector in the multiple first time sequence vector groups to obtain a cosine similarity matrix of each time sequence vector in the third time sequence vector groups, Represents the third time-series vector group m The time series vector and the first k The matrix formed by the cosine similarity of each time vector in the first time vector group. m This indicates the number of time vectors in the third time vector group. n Indicates the first k The number of time vectors in the first time vector group Represents the third time-series vector group's first... i The time series vector and the first k The first time-series vector group j Cosine similarity of time-series vectors i This indicates the time vector index of the third time vector group. j Indicates the first k The time sequence vector index of the first time sequence vector group; Summing each term in the cosine similarity matrix yields the th term. k The matching degree of the first time series vector group to the third time series vector group. , Indicates the first k The matching degree of the first time-series vector group to the third time-series vector group; the first time-series vector group whose matching degree exceeds the second threshold is determined as the first time-series vector group corresponding to the third time-series vector group. 6.The AI-based precise prediction algorithm of aeration quantity according to claim 1, wherein, if the number of first time sequence vector groups corresponding to the third time sequence vector groups is not less than 3, determining the at least one first time sequence vector group corresponding to the third time sequence vector groups as fourth time sequence vector groups comprises: calculating a first center time sequence vector of each first time sequence vector group corresponding to the third time sequence vector groups; calculating cosine similarity between first center time sequence vectors of each first time sequence vector group corresponding to the third time sequence vector group; determining, as a fourth time sequence vector group, first time sequence vector groups corresponding to two first center time sequence vectors with the largest cosine similarity.

7. The AI-based precise prediction algorithm for aeration quantity according to claim 1, wherein, If the number of second time sequence vector groups corresponding to the fourth time sequence vector group is not less than 3, determining the aeration quantity of the to-be-tested time period according to the second time sequence vector groups corresponding to the two second center time sequence vectors with the largest cosine similarity, comprises: calculating second center time sequence vectors of each second time sequence vector group corresponding to the fourth time sequence vector group; calculating cosine similarity between second center time sequence vectors of each second time sequence vector group corresponding to the fourth time sequence vector group; determining the aeration quantity of the to-be-tested time period according to the second time sequence vector groups corresponding to the two second center time sequence vectors with the largest cosine similarity. 8.The AI-based precise prediction algorithm of aeration quantity according to claim 1, wherein, determining the aeration quantity of the to-be-tested time period according to the second time sequence vector groups corresponding to the two second center time sequence vectors with the largest cosine similarity, comprises: determining, as the aeration quantity of the to-be-tested time period, the aeration quantity of the time sequence vector closest in time to the to-be-tested time period in the second time sequence vector groups corresponding to the two second center time sequence vectors with the largest cosine similarity.

Images