A digital energy air compressor station information tracing method

Abstract

The present invention relates to the field of electrical digital data processing technology, and specifically to a digital energy air compressor station information tracing method, comprising: in an information tracing system, obtaining a historical operating status data set and current operating status data of an internal cooling system of an air compressor station; determining a first correction factor and a second correction factor of a proximity distance k value when abnormal detection is performed on the current operating status data according to the historical operating status data set and the current operating status data; using the first correction factor and the second correction factor to correct a preset proximity distance k value of the current operating status data to obtain a corrected proximity distance k value of the current operating status data; based on the corrected proximity distance k value, determining whether the current operating status data is abnormal tracing data. The present invention improves the accuracy of the air compressor station information tracing result by correcting the preset proximity distance k value of the current operating status data.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data processing, and in particular to a method for tracing information of a digital energy air compressor station. Background Art

[0002] An air compression station is an air compression station. It mainly sucks in air through an air compressor and then compresses the air using a high-speed rotating rotor. In order to prevent the equipment from being damaged by excessive temperature, the compressed air is usually cooled using a cooling system to reduce the temperature of the air. However, the process of cooling the air causes the water vapor in the air to cool and condense. When the moisture content in the air is high, the moisture in the air is adsorbed in the pipeline to form water accumulation in the pipeline, which will affect the transmission efficiency of the pipeline. Therefore, the pipelines of the cooling system need to be cleaned and dredged regularly. However, the frequency of regular cleaning and dredging of pipelines often depends on experience and lacks scientific basis. Frequent cleaning may cause pipeline wear and increase maintenance costs. Infrequent cleaning may lead to reduced cooling efficiency and affect equipment performance.

[0003] In order to overcome the defects of regular cleaning and drainage pipelines, the digital energy air compressor station information traceability system is used to effectively monitor, record and analyze pipeline cleaning related data, and determine in real time whether to perform pipeline cleaning operations. When tracing abnormal data, abnormal data acquisition is often achieved through Lof (Local Outlier Factor) anomaly detection, but the pipeline cleaning results are affected by multiple dimensional factors, and the degree of influence of different dimensional factors is different. The weight of each dimensional factor in determining the distance measurement is equal in Lof anomaly detection, and the distribution changes of data points in different dimensions are not considered, which may lead to deviations in the anomaly detection results, and further make the accuracy of the pipeline abnormal status results traced by the air compressor station information traceability system low. Summary of the invention

[0004] In order to solve the above technical problem of low accuracy of the existing air compressor station information tracing results for tracing the abnormal state of the pipeline, the purpose of the present invention is to provide a digital energy air compressor station information tracing method, and the technical solution adopted is as follows:

[0005] An embodiment of the present invention provides a method for tracing information of a digital energy air compressor station, the method comprising the following steps:

[0006] In the information tracing system, the historical operating status data set and the current operating status data of the internal cooling system of the air compressor station are obtained. The operating status data is an array consisting of the air humidity, air flow rate and air pressure at the end of the air cooling pipe at the same time;

[0007] According to the historical operation status data set and the current operation status data, the degree to which the current operation status data conforms to the historical flow velocity pressure linear law is analyzed, and the first correction factor of the proximity distance k value when the current operation status data is used for abnormal detection is determined;

[0008] According to the historical operation status data set and the current operation status data, the discrete degree of the terminal air humidity of the current operation status data relative to the overall historical operation status data set is analyzed to determine the second correction factor of the proximity distance k value when the current operation status data is used for abnormal detection;

[0009] The preset proximity distance k value of the current running state data is corrected by using the first correction factor and the second correction factor to obtain the corrected proximity distance k value of the current running state data;

[0010] Based on the corrected proximity distance k value, determine whether the current operating status data is abnormal traceability data.

[0011] Further, the degree to which the current operating state data conforms to the historical flow velocity pressure linear law is analyzed based on the historical operating state data set and the current operating state data, and a first correction factor of the proximity distance k value when the current operating state data is used for abnormal detection is determined, including:

[0012] Select the air velocity and air pressure at each moment in the historical operation status data set, perform linear law analysis based on several air velocity and air pressure, and determine the linear fitting line of the historical velocity and pressure;

[0013] The air velocity and air pressure at the current moment in the current operating status data are selected, and the first correction factor of the proximity distance k value when performing abnormality detection on the current operating status data is determined according to the distribution of the data points composed of the air velocity and air pressure at the current moment relative to the linear fitting line.

[0014] Furthermore, the linear law analysis based on a plurality of air flow rates and air pressures to determine the linear fitting line of the historical flow rate pressure includes:

[0015] With the terminal air humidity as the z-axis, the air flow rate as the y-axis, and the air pressure as the x-axis, a three-dimensional coordinate system is constructed, and the historical operation status data set and the current operation status data are mapped to the three-dimensional coordinate system to construct a three-dimensional traceability model;

[0016] For the xy plane composed of air velocity and air pressure in the three-dimensional traceability model, all historical data points on the xy plane are linearly fitted to obtain a fitting straight line as the linear fitting straight line of the historical velocity pressure;

[0017] The historical data point is a data point composed of each array in the historical operation status data set.

[0018] Furthermore, the first correction factor of the proximity distance k value when performing abnormality detection on the current operating state data is determined based on the distribution of the data points composed of the air flow rate and the air pressure at the current moment relative to the linear fitting line, including:

[0019] The data point corresponding to the current running status data in the three-dimensional traceability model is recorded as the current data point; the vertical distance between the current data point and the linear fitting line, and the vertical distance between each historical data point and the linear fitting line are calculated;

[0020] Determine a first linear indicator of the current data point according to the vertical distance difference between the current data point and each historical data point and the vertical distance of the current data point;

[0021] Determine a second linear index of the current data point based on the distribution of historical data points in the local area of ​​the current data point;

[0022] In combination with the first linear index and the second linear index of the current data point, a first correction factor of the proximity distance k value when performing abnormality detection on the current running status data is determined.

[0023] Furthermore, determining the first linear index of the current data point according to the vertical distance difference between the current data point and each historical data point and the vertical distance of the current data point includes:

[0024] Calculate the absolute value of the difference between the vertical distance of the current data point and the vertical distance of each historical data point, and then calculate the cumulative value of all the absolute values ​​of the differences;

[0025] The product of the accumulated value and the vertical distance of the current data point is determined, and a negative correlation process is performed on the product to obtain a negative correlation value, and the negative correlation value is used as a first linear indicator of the current data point.

[0026] Furthermore, determining the second linear index of the current data point according to the distribution of historical data points in the local area of ​​the current data point includes:

[0027] In a two-dimensional coordinate system, determine the distance between the current data point and each historical data point, and then determine the shortest distance;

[0028] The shortest distance times the first preset value is used as the major axis value, the shortest distance times the second preset value is used as the minor axis value, the position of the current data point is used as the center of the circle, and the direction of the linear fitting line is used as the major axis direction to construct an elliptical area as the local area of ​​the current data point; wherein the first preset value is greater than the second preset value;

[0029] The number of historical data points in the local area of ​​the current data point is counted as the second linear indicator of the current data point.

[0030] Further, the step of analyzing the discrete degree of the terminal air humidity of the current operating status data relative to the overall historical operating status data set according to the historical operating status data set and the current operating status data, and determining the second correction factor of the proximity distance k value when the current operating status data is subjected to abnormality detection, includes:

[0031] Taking several terminal air humidity in the historical operation status data set as comparison data, and taking the terminal air humidity in the current operation status data as the baseline data; calculating the cumulative value of the absolute value of the difference between the baseline data and all the comparison data as the first discrete index;

[0032] The discrete conditions of the sequence composed of all the comparison data and the benchmark data are analyzed to determine the second discrete index; the second correction factor of the proximity distance k value when the current operating status data is used for abnormal detection is determined by combining the first discrete index and the second discrete index.

[0033] Furthermore, the discreteness of the sequence composed of all the comparison data and the benchmark data is analyzed to determine the second discrete index, including:

[0034] Calculate the standard deviation of the sequence consisting of all comparison data and benchmark data, and use the standard deviation as the second discrete indicator.

[0035] Furthermore, the method of using the first correction factor and the second correction factor to correct the preset proximity distance k value of the current running state data to obtain the corrected proximity distance k value of the current running state data includes:

[0036] The product of the first correction factor, the second correction factor and the preset proximity distance k value of the current running state data is calculated as the corrected proximity distance k value of the current running state data.

[0037] Furthermore, based on the corrected proximity distance k value, it is determined whether the current operating status data is abnormal tracing data, including:

[0038] The Lof anomaly detection method is used, combined with the modified proximity distance k value, to perform anomaly detection on the current data point in the three-dimensional traceability model. If the current data point is an abnormal data point, the current operating status data is determined to be abnormal traceability data. Otherwise, the current operating status data is determined not to be abnormal traceability data.

[0039] The present invention has the following beneficial effects:

[0040] The present invention provides a digital energy air compressor station information tracing method, which corrects the preset proximity distance k value of the current operating state data by analyzing the degree to which the current operating state data in the air compressor conforms to the historical flow velocity pressure linear law and the discrete degree of the terminal air humidity of the current operating state data relative to the overall historical operating state data set, and obtains the corrected proximity distance k value of the current operating state data, which considers that the operating state data of different dimensions are affected by pipeline water accumulation to different degrees, improves the accuracy of the Lof abnormal detection result, and further improves the accuracy of the air compressor station information tracing result. Based on accurate abnormal tracing data, the water accumulation in the cooling pipe of the air compressor can be judged, so that the management personnel of the information system can intelligently and real-time grasp the real-time situation in the pipeline; when the air energy output of the energy air compressor station is abnormal, the pipeline of the air compressor station can be timely checked by digitization, which is more information-based, intelligent, and accurate than regular pipeline cleaning and dredging, improves energy utilization efficiency, reduces operating costs, and provides a scientific basis for equipment maintenance and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are 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.

[0042] Figure 1 This is a flow chart of a digital energy air compressor station information tracing method according to an embodiment of the present invention;

[0043] Figure 2 4 is a flowchart for implementing step S22 in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction 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.

[0045] 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.

[0046] The application scenarios targeted by the present invention may be:

[0047] In order to prevent the air compressor temperature of the air compression station from being too high and causing damage to the equipment, a cooling system is used to cool the air when compressing it to reduce the temperature of the air during the compression process. However, this process will cause the water vapor in the air to cool and condense. When the moisture content in the air is high, the moisture in the air is adsorbed in the pipeline, forming water accumulation in the pipeline, causing pipeline corrosion and affecting the transmission efficiency of the pipeline. Among them, the cooling system of the air compression station mainly includes cooling pipelines and cooling equipment, such as hydrocondensation towers, condensate drainage pumps, and condensation pipes.

[0048] In order to avoid pipeline corrosion or affect the transmission efficiency of the pipeline, it is necessary to trace the abnormal operation status data of the pipeline in the cooling system to achieve timely cleaning and dredging of the pipeline. Specifically, this embodiment provides a digital energy air compressor station information tracing method, such as Figure 1 As shown, the following steps are included:

[0049] S1, in the information traceability system, obtain the historical operating status data set and current operating status data of the internal cooling system of the air compressor station.

[0050] Here, when the digital energy air compressor station is in operation, various types of sensors are usually used to monitor the operating status of the air compressor cooling system in real time. Because water accumulation in the cooling pipe of the cooling system may affect the pressure, flow rate and air humidity of the air in the pipe, in order to monitor the abnormal state of water accumulation in the cooling system pipe, an air humidity sensor, air pressure sensor and anemometer are set in the cooling system. After the data is collected, it is transmitted to the data center or cloud backend management system through the network to realize the traceability of the digital energy air compressor station information, and then the cooling system is controlled in real time based on the traceability data to clean and dredge the water in the pipe.

[0051] Specifically, the information traceability system first obtains the historical operating status data set of the cooling system inside the air compressor station from the database through the sensors installed in the cooling system, and then collects the current operating status data of the cooling system inside the air compressor station in real time. Among them, the historical operating status data set contains the operating status data at each moment in the historical preset period. The collection frequency of the operating status data can be set to collect once every second, and the collection period is 30 minutes. The operating status data specifically refers to the array composed of the air humidity, air flow rate and air pressure at the end of the air cooling pipe at the same time, and the current operating status data refers to the array composed of the air humidity, air flow rate and air pressure at the end of the air cooling pipe at the current moment.

[0052] The sensors here specifically include air humidity sensors, air pressure sensors, and anemometers. The implementer can set the collection frequency and collection period of historical operating status data according to the actual situation.

[0053] It is worth noting that the historical operating status data set is collected under the condition of ensuring that there is no water accumulation in the pipes of the cooling system. It can be used to build a three-dimensional traceability model. Data analysis is performed through this model to determine in real time whether there is water accumulation in the pipes corresponding to the current real-time collected operating status data. If there is water accumulation, the pipes need to be cleaned or unblocked to avoid corrosion damage to the equipment in the air compressor station.

[0054] So far, this embodiment has obtained the basic data for information traceability analysis, the historical operation status data set and the current operation status data.

[0055] S2, analyzing the degree to which the current operating status data conforms to the historical flow velocity pressure linear law based on the historical operating status data set and the current operating status data, and determining the first correction factor of the proximity distance k value when performing abnormality detection on the current operating status data.

[0056] It should be noted that the existing Lof anomaly detection algorithm is often used to trace abnormal operating status data, and the proximity distance k value in the existing Lof anomaly detection algorithm is determined by using data with equal weights in different dimensions, without considering the different distribution changes of data points in different dimensions. In the air change scenario of tracing the abnormal information of the cooling pipeline of the air compressor station, when the air passes through the cooling pipeline, the flow rate of the air is proportional to the air pressure, and the air flow rate and pressure are linearly distributed, so the weight of the current operating status data that conforms to the linear law of flow rate and pressure should be increased; in addition, the condensation and accumulation of pipeline air has the greatest impact on the humidity of the air, because the accumulation of water in the pipeline will make the air humidity in the cooling pipeline higher, and it should have a greater influence weight on the air humidity. Therefore, it is necessary to correct the preset proximity distance k value. First, based on the historical operating status data set and the current operating status data, the degree to which the current operating status data conforms to the historical flow rate and pressure linear law is analyzed, and the first correction factor of the proximity distance k value when the current operating status data is detected for abnormality is determined.

[0057] The above step S2 can be implemented by the following steps S21 to S22 (not shown):

[0058] S21, selecting the air flow rate and air pressure at each moment in the historical operation status data set, performing linear law analysis based on a plurality of air flow rates and air pressures, and determining a linear fitting line of the historical flow rate and pressure.

[0059] In order to analyze the linear distribution law of flow velocity and pressure, the air velocity and air pressure at each moment in the historical operating status data set are first obtained, and the distribution law of the obtained historical air velocity and pressure data is analyzed using a linear fitting algorithm, so as to obtain a more standard data distribution corresponding to the linear law of flow velocity and pressure, that is, to determine the linear fitting line of the historical flow velocity and pressure.

[0060] The above step S21 can be implemented by the following steps S211 to S212 (not shown):

[0061] S211, construct a three-dimensional coordinate system with the terminal air humidity as the z-axis, the air flow rate as the y-axis, and the air pressure as the x-axis, map the historical operation status data set and the current operation status data to the three-dimensional coordinate system, and construct a three-dimensional traceability model.

[0062] In this embodiment, when the cooling pipe is normal and there is no water accumulation, the changes in the air pressure, air flow rate and air humidity in the pipe are small, and its historical data distribution is relatively concentrated. Therefore, when analyzing the linear relationship between air pressure and air flow rate, the air humidity at the end of the cooling pipe is used as the z-axis, the air flow rate is used as the y-axis, and the air pressure is used as the x-axis to construct a three-dimensional space coordinate system, and the collected historical operation status data set and the current operation status data are mapped to the three-dimensional coordinate system to form a corresponding number of historical data points and current data points, and the three-dimensional space coordinate system at this time is used as a three-dimensional traceability model. Among them, the historical data point is the data point composed of each array in the historical operation status data set, and the current data point refers to the data point composed of the three-dimensional data of the current operation status data, that is, the data point composed of the array at the current moment.

[0063] S212, for the xy plane composed of air velocity and air pressure in the three-dimensional traceability model, linear fitting is performed on all historical data points on the xy plane to obtain a fitting straight line as the linear fitting straight line of the historical flow velocity pressure.

[0064] In this embodiment, when compressed air flows in the pipeline, it is affected by the air pressure and flows from one end of the pipeline to the other end of the pipeline. The greater the pressure of the compressed air, the faster the air flow rate in the pipeline, that is, the air flow rate in the cooling pipeline is linearly related to the air pressure. The distribution of historical data points in the three-dimensional traceability model on the xy-axis plane composed of air pressure and air flow rate should change linearly. Then, when the current data point satisfies the linear relationship in the pressure-flow rate plane, that is, the current data point of the adjacent distance k value belongs to the linearly distributed data point, which further indicates that the current data point corresponding to the adjacent distance k value should be selected.

[0065] Specifically, the least square method is used to perform linear fitting on all historical data points on the xy plane to obtain a fitting straight line, which is used as the linear fitting straight line of the historical flow velocity pressure. The implementation process of the least square method is prior art and is not within the scope of protection of the present invention, and will not be elaborated in detail here; the independent variable of the fitting straight line is the air pressure, and the dependent variable is the air velocity.

[0066] S22, select the air flow rate and air pressure at the current moment in the current operating status data, and determine the first correction factor of the proximity distance k value when performing abnormality detection on the current operating status data based on the distribution of the data points composed of the air flow rate and air pressure at the current moment relative to the linear fitting line.

[0067] Here, the first correction factor is one of the key factors for correcting the proximity k value of the current data point. The larger the first correction factor is, the greater the degree to which the preset proximity k value is corrected, and the larger the subsequently determined corrected proximity k value will be.

[0068] The above step S22 can be performed by Figure 2 The steps shown achieve:

[0069] S221, record the data point corresponding to the current operating status data in the three-dimensional traceability model as the current data point; calculate the vertical distance between the current data point and the linear fitting line, and the vertical distance between each historical data point and the linear fitting line.

[0070] In this embodiment, the air flow rate and air pressure at the current moment in the current operating status data are selected, that is, the position of the current data point in the three-dimensional traceability model is determined, so as to facilitate the subsequent analysis of the distance between the current data point and the linear fitting line formed by the historical data points on the flow rate pressure plane. The closer the current data point is to the linear fitting line, the greater the degree to which the current data point conforms to the linear law of the historical flow rate pressure, and the larger the first correction factor.

[0071] Specifically, on the velocity pressure plane, a line segment perpendicular to the linear fitting line is drawn through the current data point, and the length of the selected segment is determined as the vertical distance of the current data point; at the same time, the vertical distance of each historical data point is determined. The vertical distance is used to measure the degree to which the data point conforms to the law of velocity pressure change. The calculation process of determining the vertical distance between two points based on position is a prior art and is not within the scope of protection of the present invention, and will not be repeated here.

[0072] S222, determining a first linear index of the current data point according to the vertical distance difference between the current data point and each historical data point and the vertical distance of the current data point.

[0073] Here, the first linear indicator can characterize the distance between the current data point and the linear fitting line. The closer the current data point is to the linear fitting line corresponding to the historical operating status data set, the greater the degree to which the current data point conforms to the historical flow velocity pressure linear data distribution law, and the better the linear relationship of the current data point.

[0074] The above step S222 can be implemented by the following steps:

[0075] First, the absolute value of the difference between the vertical distance of the current data point and the vertical distance of each historical data point is calculated, and then the cumulative value of all the absolute values ​​of the differences is calculated.

[0076] Secondly, the product of the accumulated value and the vertical distance of the current data point is determined, and the product is negatively correlated to obtain a negative correlation value, and the negative correlation value is used as the first linear indicator of the current data point.

[0077] It should be noted that, relative to the vertical distance of all historical data points, the smaller the vertical distance of the current data point, the smaller the vertical distance difference between the two, the more the current data point conforms to the linear distribution law of velocity pressure, and the stronger the linear relationship of the current data point; the smaller the vertical distance of the current data point, the greater the possibility that the current data point is on the linear fitting line, and the stronger the linear relationship of the current data point. Therefore, the vertical distance difference and vertical distance of the current data point are both negatively correlated with the first linear indicator, so it is necessary to perform negative correlation processing on the product to obtain a negative correlation value, and use the negative correlation value as the first linear indicator of the current data point. The implementation method of the negative correlation processing can be an exponential function with a natural constant as the base, that is, exp (-), which is not specifically limited.

[0078] S223, determining a second linear index of the current data point according to the distribution of historical data points in the local area of ​​the current data point.

[0079] Here, the second linear index is to measure the distribution density of historical data points in the local area of ​​the current data point. The higher the distribution density of historical data points in the local area, the higher the similarity, stability and flow characteristics of the data points in the local area of ​​the current data point, which can make the relationship between pressure and flow velocity more linear both at the statistical and physical levels.

[0080] The above step S223 can be implemented by the following steps:

[0081] First, in the two-dimensional coordinate system, the distance between the current data point and each historical data point is determined, and then the shortest distance is determined.

[0082] Secondly, the shortest distance of the first preset multiple is taken as the major axis value, the shortest distance of the second preset multiple is taken as the minor axis value, the position of the current data point is taken as the center of the circle, and the direction of the linear fitting line is taken as the major axis direction, and an elliptical area is constructed as the local area of ​​the current data point. Among them, the first preset multiple is greater than the second preset multiple, the first preset multiple takes an empirical value of five times, and the second preset multiple takes an empirical value of two times.

[0083] Then, the number of historical data points in the local area of ​​the current data point is counted as the second linear indicator of the current data point.

[0084] It should be noted that there is no limitation on the process of determining the local area of ​​the current data point, and the implementer can adaptively set a more suitable local area for the current data point according to the specific data distribution situation; the higher the data point distribution density in the local area where the current data point is located, the more the current data point conforms to the linear law of the flow velocity pressure change distribution in the cooling pipe, that is, the better the linear relationship of the current data point, the greater the possibility that the current data point is selected, and the greater the possibility that it is located within the range of the adjacent distance k value.

[0085] S224, combining the first linear index and the second linear index of the current data point to determine a first correction factor of the proximity distance k value when performing abnormality detection on the current running status data.

[0086] In this embodiment, the first linear index, the second linear index and the first correction factor are positively correlated. The larger the first linear index and the second linear index are, the larger the first correction factor is. Therefore, the product or addition of the first linear index and the second linear index can be used as the first correction factor.

[0087] Specifically, the product of the first linear index and the second linear index is calculated, and the product of the two linear indexes is normalized using a linear normalization function to obtain a normalized value, which is used as a first correction factor of the proximity distance k value when performing abnormality detection on the current operating status data.

[0088] As a strength, the calculation formula for the first correction factor of the current data point can be:

[0089] ; In the formula, represents the first correction factor of the current data point, z represents the current data point, and exp represents the exponential function with the natural constant as the base. represents the vertical distance of the current data point, X represents the number of historical data points, and n represents the sequence number of the historical data point. represents the vertical distance of the nth data point, Represents the first linear indicator of the current data point, Indicates the number of historical data points in the local area of ​​the current data point, It also represents the second linear index of the current data point. exp represents an exponential function with a natural constant as the base. exp(-) can be used to achieve normalization processing of the negative correlation of distance.

[0090] It should be noted that the first correction factor is determined from two aspects: one is the distance between the current data point and the pressure-flow velocity linear fitting line, and the other is the distribution of historical data points in the local area of ​​the current data point. The numerical accuracy of the first correction factor determined in this way is higher, which helps to improve the accuracy of the information traceability results of the digital energy air compressor station.

[0091] So far, this embodiment obtains the first correction factor of the proximity distance k value when performing abnormality detection on the current running status data.

[0092] S3, analyzing the discrete degree of the terminal air humidity of the current operating status data relative to the overall historical operating status data set according to the historical operating status data set and the current operating status data, and determining a second correction factor of the proximity distance k value when performing abnormality detection on the current operating status data.

[0093] It should be noted that in the cooling system of the air compressor station, when the compressed air is cooled in the pipeline, the air is compressed, making the gas molecules in the air closer together, which will also make the water molecules in the air easier to store in the air. Condensation occurs due to the cooling of the pipeline, forming condensed water. The condensed water gathers in the cooling pipeline, making the pipeline moist, thereby accelerating the change of air humidity in the pipeline. Therefore, when using the Lof local outlier factor to detect abnormal data points, in the calculation process of the proximity distance k value of the current data point, the air humidity can better reflect the abnormal change of the data point for the abnormal change of the air, so the air humidity has a greater weight on the proximity distance k value.

[0094] In this embodiment, the current data point differs from the historical data points in the local area in the dimension of air humidity. If the difference between the current data point and the historical data points in the local area is large, it means that the weight of the air humidity of the current data point is greater. Because the air humidity changes in the cooling pipe are relatively different, that is, the distribution of the air humidity in the three-dimensional traceability model is relatively scattered, indicating that the air humidity in the cooling pipe changes greatly, and the air humidity changes in the cooling pipe are relatively different. Therefore, the weight of the air humidity is greater in the calculation process of the proximity distance k value.

[0095] The above step S3 can be implemented through steps S31 to S32 (not shown in the figure):

[0096] S31, taking several terminal air humidity in the historical operation status data as comparison data, and taking the terminal air humidity in the current operation status data as reference data; calculating the cumulative value of the absolute value of the difference between the reference data and all the comparison data as the first discrete index.

[0097] Here, the first discrete index represents the degree of outlier of the terminal air humidity at the current moment relative to all historical terminal air humidities in the local area. The greater the degree of outlier, the greater the change in air humidity in the cooling duct, the greater the first discrete index, and thus the greater the second correction factor.

[0098] In this embodiment, in order to facilitate understanding of the solution, multiple terminal air humidity is selected from the historical operation status data set as comparison data, and the comparison data is used for comparison with the benchmark data, and the terminal air humidity in the current operation status data is used as the benchmark data. When selecting the comparison data, as an example, a circle is extended outward with the current data point as the center in the three-dimensional traceability model, and the circle is terminated when the density of data points in the circle is the largest, and the historical data point in the circle at this time is used as the comparison data point of the current data point.

[0099] S32, analyzing the discreteness of the sequence composed of all the comparison data and the benchmark data to determine a second discrete index; combining the first discrete index and the second discrete index to determine a second correction factor of the proximity distance k value when performing abnormality detection on the current operating status data.

[0100] In this embodiment, the first discrete index, the second discrete index and the second correction factor are positively correlated. The larger the first discrete index and the second discrete index, the larger the second correction factor. Therefore, the product or sum of the two discrete indexes can be used as the second correction factor.

[0101] Specifically, the standard deviation of the sequence consisting of all comparison data and benchmark data is calculated as the second discrete index; the product of the first discrete index and the second discrete index is calculated, and the product is normalized using a normalization function to obtain a normalized value, and the normalized value is used as the second correction factor of the proximity distance k value when performing anomaly detection on the current operating status data.

[0102] The calculation formula of the second correction factor of the proximity distance k value when the current running status data is used for abnormal detection can be:

[0103] ; In the formula, Indicates the second correction factor of the proximity distance k value when performing anomaly detection on the current operating status data. represents the normalization function, c represents the sequence number of the comparison data, C represents the number of comparison data, Represents the benchmark data, represents the cth comparison data, represents the absolute value function, represents the first discrete index, Represents the standard deviation of the sequence consisting of all comparison data and benchmark data. It also represents the second discrete index.

[0104] In the calculation formula of the second correction factor, when the first discrete index When the value is larger, it means that the difference between the air humidity of the current data point and the historical data points in the local area is greater, indicating that the difference between the air humidity of the current data point and the air humidity of the historical data points is greater; when the second discrete index The larger it is, the more dispersed the distribution of the current data point and all historical data points on the air humidity axis is, which means that the difference between the current data point on the air humidity axis is large; when the second correction factor is larger, it means that the weight of the air humidity is greater when the current data point is calculating the proximity distance k value.

[0105] So far, this embodiment has obtained the second correction factor of the proximity distance k value when performing abnormality detection on the current running status data.

[0106] S4, using the first correction factor and the second correction factor to correct the preset proximity distance k value of the current running state data, to obtain a corrected proximity distance k value of the current running state data.

[0107] In this embodiment, the preset proximity k value of the current operating status data, that is, the proximity k value of the current data point when performing local anomaly detection on the three-dimensional traceability model, the initially set proximity k value can be determined during the implementation of the Lof anomaly detection algorithm, which will not be repeated here. Since the preset proximity k value does not take into account the different degrees of influence of the operating data of different dimensions of the actual cooling system on the abnormality of water accumulation in the pipeline, the first correction factor and the second correction factor are used to correct the preset proximity k value, so as to obtain the corrected proximity k value, and then judge the abnormal traceability data of the current data point.

[0108] Specifically, the product of the first correction factor, the second correction factor, and the preset proximity distance k value of the current running state data is calculated as the corrected proximity distance k value of the current running state data.

[0109] It should be noted that choosing a suitable proximity distance k value is crucial to the performance of the Lof algorithm. The reason is that if the proximity distance k value is too small, noise points may be mistaken for outliers, because the neighbors of each point may be just some random neighboring points rather than a real group structure; if the k value is too large, all points may be regarded as normal points, ignoring the detection of small groups or outliers.

[0110] So far, this embodiment has obtained the modified proximity distance k value for anomaly detection.

[0111] S5, based on the corrected proximity distance k value, determine whether the current operating status data is abnormal tracing data.

[0112] Specifically, the Lof anomaly detection method is used in combination with the modified proximity distance k value to perform anomaly detection on the current data point in the three-dimensional traceability model. If the current data point is an abnormal data point, the current operating status data is determined to be abnormal tracing data. Otherwise, the current operating status data is determined not to be abnormal tracing data.

[0113] In this embodiment, when the operating status data at the current moment is abnormal traceability data, it may indicate that there is water accumulation in the cooling pipes of the cooling system of the digital energy air compressor station at the current moment, and the water accumulation is relatively serious, indicating that the possibility of water accumulation in the pipes at this moment is relatively high, and the digital intelligent platform will perform corresponding system early warning prompts after information tracing; when the operating status data at the current moment is not abnormal traceability data, it indicates that the water accumulation in the cooling pipes of the digital energy air compressor station at the current moment is general, and frequent pipe cleaning is not required, and the system does not give special prompts.

[0114] This concludes the present embodiment.

[0115] The present invention provides a digital energy air compressor station information tracing method, which analyzes the degree to which the current operating status data in the air compressor conforms to the historical flow rate pressure linear law and the terminal air humidity of the current operating status data relative to the overall discrete degree of the historical operating status data set, and corrects the preset proximity distance k value of the current operating status data to obtain the corrected proximity distance k value of the current operating status data, which considers that the operating status data of different dimensions are affected by pipeline water accumulation to different degrees, improves the accuracy of the Lof abnormal detection result, and further improves the accuracy of the air compressor station information tracing result. Based on accurate abnormal tracing data, the water accumulation in the cooling pipeline of the air compressor can be judged, so that the management personnel of the information system can intelligently and real-time grasp the real-time situation in the pipeline; when the air energy output of the energy air compressor station is abnormal, the pipeline of the air compressor station can be timely checked by digitization, which is more information-based, intelligent, and accurate than regular pipeline cleaning and dredging, improves energy utilization efficiency, reduces operating costs, and provides a scientific basis for equipment maintenance and optimization.

[0116] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A digital energy air compressor station information tracing method, characterized in that: The following steps are involved: In the information tracing system, the historical operating status data set and the current operating status data of the internal cooling system of the air compressor station are obtained. The operating status data is an array consisting of the air humidity, air flow rate and air pressure at the end of the air cooling pipe at the same time; According to the historical operation status data set and the current operation status data, the degree to which the current operation status data conforms to the historical flow velocity pressure linear law is analyzed, and the first correction factor of the proximity distance k value when the current operation status data is used for abnormal detection is determined; According to the historical operation status data set and the current operation status data, the discrete degree of the terminal air humidity of the current operation status data relative to the overall historical operation status data set is analyzed, and the second correction factor of the proximity distance k value when the current operation status data is used for abnormal detection is determined; The preset proximity distance k value of the current running state data is corrected by using the first correction factor and the second correction factor to obtain the corrected proximity distance k value of the current running state data; Based on the corrected proximity distance k value, determine whether the current operating status data is abnormal traceability data; According to the distribution of the data points composed of the air velocity and the air pressure at the current moment relative to the linear fitting line, the first correction factor of the proximity distance k value when the current operating state data is used for abnormal detection is determined, including: The data point corresponding to the current running status data in the three-dimensional traceability model is recorded as the current data point; the vertical distance between the current data point and the linear fitting line, and the vertical distance between each historical data point and the linear fitting line are calculated; Determine a first linear indicator of the current data point according to the vertical distance difference between the current data point and each historical data point and the vertical distance of the current data point; Determine a second linear index of the current data point based on the distribution of historical data points in the local area of ​​the current data point; Combining the first linear index and the second linear index of the current data point, determining a first correction factor of the proximity distance k value when performing abnormality detection on the current operating state data; Determining the first linear index of the current data point according to the vertical distance difference between the current data point and each historical data point and the vertical distance of the current data point includes: Calculate the absolute value of the difference between the vertical distance of the current data point and the vertical distance of each historical data point, and then calculate the cumulative value of all the absolute values ​​of the differences; The product of the accumulated value and the vertical distance of the current data point is determined, and a negative correlation process is performed on the product to obtain a negative correlation value, and the negative correlation value is used as a first linear indicator of the current data point.

2. According to claim 1, a digital energy air compressor station information tracing method is characterized in that: The method of analyzing the degree to which the current operating state data conforms to the historical flow velocity pressure linear law according to the historical operating state data set and the current operating state data, and determining the first correction factor of the proximity distance k value when the current operating state data is used for abnormal detection, includes: Select the air velocity and air pressure at each moment in the historical operation status data set, perform linear law analysis based on several air velocity and air pressure, and determine the linear fitting line of the historical velocity and pressure; The air velocity and air pressure at the current moment in the current operating status data are selected, and the first correction factor of the proximity distance k value when performing abnormality detection on the current operating status data is determined according to the distribution of the data points composed of the air velocity and air pressure at the current moment relative to the linear fitting line.

3. A digital energy air compressor station information tracing method according to claim 2, characterized in that: The linear law analysis based on a plurality of air flow rates and air pressures is performed to determine the linear fitting line of the historical flow rate pressure, including: With the terminal air humidity as the z-axis, the air flow rate as the y-axis, and the air pressure as the x-axis, a three-dimensional coordinate system is constructed, and the historical operation status data set and the current operation status data are mapped to the three-dimensional coordinate system to construct a three-dimensional traceability model; For the xy plane composed of air velocity and air pressure in the three-dimensional traceability model, all historical data points on the xy plane are linearly fitted to obtain a fitting straight line as the linear fitting straight line of the historical velocity pressure; The historical data point is a data point composed of each array in the historical operation status data set.

4. According to claim 1, a digital energy air compressor station information tracing method is characterized in that: Determining the second linear index of the current data point according to the distribution of historical data points in the local area of ​​the current data point includes: In the two-dimensional coordinate system, determine the distance between the current data point and each historical data point, and then determine the shortest distance; The shortest distance times the first preset value is used as the major axis value, the shortest distance times the second preset value is used as the minor axis value, the position of the current data point is used as the center of the circle, and the direction of the linear fitting line is used as the major axis direction to construct an elliptical area as the local area of ​​the current data point; wherein the first preset value is greater than the second preset value; The number of historical data points in the local area of ​​the current data point is counted as the second linear indicator of the current data point.

5. According to claim 1, a digital energy air compressor station information tracing method is characterized in that: The method of analyzing the discrete degree of the terminal air humidity of the current operating state data relative to the overall historical operating state data set according to the historical operating state data set and the current operating state data, and determining the second correction factor of the proximity distance k value when the current operating state data is subjected to abnormality detection, includes: Taking several terminal air humidity in the historical operation status data set as comparison data, and taking the terminal air humidity in the current operation status data as the baseline data; calculating the cumulative value of the absolute value of the difference between the baseline data and all the comparison data as the first discrete index; The discrete conditions of the sequence composed of all the comparison data and the benchmark data are analyzed to determine the second discrete index; the second correction factor of the proximity distance k value when the current operating status data is used for abnormal detection is determined by combining the first discrete index and the second discrete index.

6. A digital energy air compressor station information tracing method according to claim 5, characterized in that: Analyze the discreteness of the sequence composed of all comparative data and benchmark data, and determine the second discrete index, including: Calculate the standard deviation of the sequence consisting of all comparison data and benchmark data, and use the standard deviation as the second discrete indicator.

7. A digital energy air compressor station information tracing method according to claim 1, characterized in that: The method of correcting the preset proximity distance k value of the current running state data by using the first correction factor and the second correction factor to obtain the corrected proximity distance k value of the current running state data includes: The product of the first correction factor, the second correction factor and the preset proximity distance k value of the current running state data is calculated as the correction proximity distance k value of the current running state data.

8. A digital energy air compressor station information tracing method according to claim 1, characterized in that: Based on the corrected proximity distance k value, determine whether the current operating status data is abnormal traceability data, including: The Lof anomaly detection method is used, combined with the modified proximity distance k value, to perform anomaly detection on the current data point in the three-dimensional traceability model. If the current data point is an abnormal data point, the current operating status data is determined to be abnormal traceability data. Otherwise, the current operating status data is determined not to be abnormal traceability data.

Images