A method and system for detecting anomalies in combustible gas monitoring data

Abstract

This invention relates to the field of data processing technology, specifically to a method and system for detecting anomalies in combustible gas monitoring data. The method includes: acquiring combustible gas emission data curves for each monitoring point; dividing the data curves of each monitoring point into several curve segments; identifying the standard curve segment of the combustible gas emission data curve for each monitoring point; determining the degree of data anomaly for each curve segment based on the difference between each curve segment and the standard curve segment of that monitoring point; combining the overlap of time distributions between curve segments from different monitoring points to obtain several curve segment groups; calculating the probability of wind occurrence; and finally determining all actual abnormal curve segments of the combustible gas emission data curve. This invention, by dynamically dividing the data curves and identifying standard curve segments, and analyzing data anomalies caused by wind, provides detailed analysis and differentiation, avoiding resource waste and unnecessary intervention due to misjudgment.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for detecting anomalies in combustible gas monitoring data. Background Technology

[0002] Combustible gases, such as hydrogen, methane, butane, ethylene, acetylene, natural gas, and liquefied petroleum gas, possess flammability or explosiveness under conditions of heat or impact. When combustible gases are stored in factory warehouses or buried underground via pipelines, the limited ventilation in these confined spaces can cause the concentration of combustible gases to reach the flammability and explosive limits during a leak, making an explosion highly likely and resulting in severe economic losses and personnel safety accidents.

[0003] In pressure regulating stations, valve wells, and underground pipeline networks, multi-sensor IoT terminals are deployed at key nodes to detect and locate gas leaks using flow anomalies based on graphical data structures. However, in working environments such as mines and factories, factors such as temperature changes, humidity, corrosive gases (poisoning), and aging significantly interfere with the data collected by the sensors. Existing anomaly data is usually scarce, and data imbalance is a serious problem. Most detection models perform well on normal data but cannot effectively handle data anomalies related to combustible gases under special working conditions. Summary of the Invention

[0004] The present invention provides a method and system for detecting abnormal combustible gas monitoring data, which adopts the following technical solution:

[0005] One embodiment of the present invention provides a method for detecting anomalies in combustible gas monitoring data, the method comprising the following steps:

[0006] Obtain the combustible gas emission data curve for each monitoring point;

[0007] Based on the combustible gas emission data curves of all monitoring points and the preset reference range for each data point, the combustible gas emission data curve of each monitoring point is divided into several curve segments;

[0008] Based on the standard deviation of combustible gas emission data for all data points in each curve segment of the combustible gas emission data curve at each monitoring point, and the DTW distance between the curve segment and the combustible gas emission data, the standard curve segment of the combustible gas emission data curve at each monitoring point is obtained; based on the difference between each curve segment of the combustible gas emission data curve at each monitoring point and its standard curve segment, the degree of data anomaly of each curve segment of the combustible gas emission data curve at each monitoring point is obtained.

[0009] Based on the overlap of all curve segments and their time distributions in the combustible gas emission data curves of all monitoring points, several curve segment groups are obtained for the combustible gas emission data curves of all monitoring points. Based on the degree of data anomaly and the duration of all curve segments within each curve segment group, the probability of wind in each curve segment group is obtained. Based on the probability of wind in each curve segment group and the average degree of data anomaly of all curve segments within each curve segment group, all actual abnormal curve segments of the combustible gas emission data curves of all monitoring points are obtained.

[0010] Furthermore, the specific steps involved in dividing the combustible gas emission data curve of each monitoring point into several curve segments based on the combustible gas emission data curves of all monitoring points and the preset reference range for each data point are as follows:

[0011] The data curve segment of combustible gas emission with the left adjacent reference time for each data point is preset and recorded as the reference range for each data point;

[0012] The first data point of the combustible gas emission data curve for each monitoring point is recorded as the first segment point of the combustible gas emission data curve for each monitoring point.

[0013] Based on the combustible gas emission data curves of all monitoring points and the reference range of each data point, the segmented evaluation index of each data point of the combustible gas emission data curve of each monitoring point is obtained;

[0014] When the segmentation evaluation index of any data point in the combustible gas emission data curve of any monitoring point is greater than the preset judgment threshold, that data point in the combustible gas emission data curve of that monitoring point is taken as the segmentation point of the combustible gas emission data curve of that monitoring point.

[0015] Based on several segment points of the combustible gas emission data curves of all monitoring points, the combustible gas emission data curve of each monitoring point is divided into several curve segments.

[0016] Furthermore, based on the combustible gas emission data curves of all monitoring points and the reference range for each data point, the specific calculation formula for the segmented evaluation index of each data point of the combustible gas emission data curve of each monitoring point is as follows:

[0017]

[0018] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point Segmented evaluation indicators for each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point The maximum value of combustible gas emissions between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point The minimum combustible gas emission of each data point between its previous segment point and all other data points. Indicates the first The first curve of combustible gas emission data at each monitoring point The number of all data points between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point Combustible gas emissions at each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point Among all data points between the nth data point and its previous segmentation point, the nth data point is... Combustible gas emissions at each data point; This indicates the total number of monitoring points; Indicates except the first All other monitoring points besides the first monitoring point were at the [number]th monitoring point. The number of monitoring points with segmented points within the reference range of each data point; It is an absolute value function.

[0019] Furthermore, the specific steps for obtaining the standard curve segment of the combustible gas emission data curve for each monitoring point based on the standard deviation of the combustible gas emission data of all data points in each curve segment of the combustible gas emission data curve for each monitoring point are as follows:

[0020] Based on the standard deviation of the combustible gas emission data of all data points in each curve segment of the combustible gas emission data curve of each monitoring point, and the DTW distance between the curve segment and the combustible gas emission data, the standard evaluation of each curve segment of the combustible gas emission data curve of each monitoring point is obtained.

[0021] Among all the curve segments of the combustible gas emission data curve at any monitoring point, the curve segment with the largest standard evaluation is taken as the standard curve segment of the combustible gas emission data curve at that monitoring point.

[0022] Furthermore, the specific calculation formula for the standard evaluation of each curve segment of the combustible gas emission data curve obtained for each monitoring point is as follows:

[0023]

[0024] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point Standard evaluation of each curve segment; Indicates the first The first curve of combustible gas emission data at each monitoring point The standard deviation of combustible gas emission data for all data points in each curve segment; Indicates the first The combustible gas emission data curve of the first monitoring point and its corresponding first monitoring point The DTW distance of each curve segment.

[0025] Furthermore, the specific formula for calculating the degree of data anomaly of each curve segment of the combustible gas emission data curve for each monitoring point, based on the difference between each curve segment and its standard curve segment, is as follows:

[0026]

[0027] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point The degree of data anomaly in each curve segment; Indicates the first The first curve of combustible gas emission data at each monitoring point The mean of all combustible gas emission data for each curve segment; Indicates the first The mean of all combustible gas emission data in the standard curve segment of the combustible gas emission data curve for each monitoring point; Indicates the first Standard evaluation of the standard curve segment of the standard curve segment of the combustible gas emission data curve for each monitoring point; Indicates the first The first curve of combustible gas emission data at each monitoring point Standard evaluation of each curve segment; It is an absolute value function.

[0028] Furthermore, the specific steps for obtaining several sets of curve segments of the combustible gas emission data curves for all monitoring points based on the overlap of all curve segments and their time distributions are as follows:

[0029] The time range corresponding to any curve segment of the combustible gas emission data curve of any monitoring point, and the reference range of the first and last data points of that curve segment, is all the curve segments of the combustible gas emission data curve of each other monitoring point. This is recorded as the reference curve segment of the combustible gas emission data curve of that monitoring point in each other monitoring point.

[0030] For the combustible gas emission data curve segment of the monitoring point, when there are multiple reference curve segments at any other monitoring point, the reference curve segment with the most time axis overlap is selected.

[0031] Calculate the total time overlap length of each curve segment of the combustible gas emission data curve at each monitoring point with all its reference curve segments. Each curve segment and all its reference curve segments constitute a curve segment combination, and each curve segment combination corresponds to a total time overlap length. If curve segments overlap between curve segment combinations, the curve segment combination with the largest total time overlap length is taken as a curve segment group.

[0032] Furthermore, the specific calculation formula for the wind probability of each curve segment group, based on the degree of data anomaly and time length of all curve segments within each curve segment group, is as follows:

[0033]

[0034] In the formula, Indicates the first The probability of wind in each curve segment group; Indicates the first The sum of the differences in time length between any two curve segments in a group of curve segments; Indicates the first The sum of the differences in the degree of data anomaly between any two curve segments in a group of curve segments; It is an exponential function with the natural constant as its base.

[0035] Furthermore, the specific steps for obtaining all actual abnormal curve segments of the combustible gas emission data curves for all monitoring points based on the wind probability of each curve segment group and the average degree of data anomaly of all curve segments within each curve segment group are as follows:

[0036] When the probability of wind in any curve segment group is less than a preset judgment threshold and the average value of the data anomaly degree of all curve segments in the curve segment group is greater than a preset anomaly threshold, all curve segments in the curve segment group are considered as actual abnormal curve segments.

[0037] The present invention also proposes a combustible gas monitoring data anomaly detection system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned combustible gas monitoring data anomaly detection method.

[0038] The beneficial effects of the technical solution of the present invention are:

[0039] In this embodiment of the invention, by segmenting the data curve, the changes in combustible gas emissions over different time periods can be analyzed more precisely, revealing subtle fluctuations and trends in emissions. Next, the volatility and correlation of combustible gas emission data within each curve segment are evaluated. Statistical methods such as standard deviation and DTW distance are used to quantify the intrinsic correlation and dispersion between the curve segment and the standard curve segment. Furthermore, combining curve segments from different monitoring points allows for a comprehensive review of emissions from multiple monitoring points within the same time period, revealing regional emission characteristics and potential environmental problems. Further evaluation of the correlation between the curve segment group and a specific situation (such as emission changes caused by wind) ultimately enables accurate identification of actual abnormal curve segments among all monitoring points, improving the accuracy and reliability of combustible gas emission anomaly detection. Thus, this invention, through dynamic segmentation of data curves and identification of standard curve segments, and analysis of data anomalies caused by wind, provides detailed analysis and differentiation, avoiding resource waste and unnecessary intervention due to misjudgment. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of the steps in the method for detecting abnormal combustible gas monitoring data according to the present invention. Detailed Implementation

[0042] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a combustible gas monitoring data anomaly detection method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0043] Unless otherwise defined, 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 pertains.

[0044] The following description, in conjunction with the accompanying drawings, details the specific scheme of the combustible gas monitoring data anomaly detection method and system provided by the present invention.

[0045] Please see Figure 1 The diagram illustrates a flowchart of a method for detecting anomalies in combustible gas monitoring data according to an embodiment of the present invention. The method includes the following steps:

[0046] Step S001: Obtain the combustible gas emission data curve for each monitoring point.

[0047] When monitoring combustible gas emissions, a stable flow field is required for measuring smoke and velocity. Unstable flow fields, with their swirling and eddy currents, can lead to uneven smoke distribution and drastic velocity variations. Therefore, selecting a cross-section with a relatively stable flue gas flow field for monitoring is crucial. Direct monitoring at the combustible gas emission pipe outlet may be affected by instability in the internal flow field of the pipe, resulting in inaccurate data. Monitoring at multiple locations in the surrounding area can average out the instability that may occur at a single location, improving data accuracy. Furthermore, multi-point monitoring facilitates the analysis of pollutant diffusion paths and concentration distribution, thereby enabling the tracing of the pollution source's location and emission characteristics.

[0048] Several monitoring points were manually selected around the emission pipeline, and the combustible gas emission data curves for each monitoring point were obtained at five-minute intervals.

[0049] Step S002: Based on the combustible gas emission data curves of all monitoring points and the preset reference range for each data point, divide the combustible gas emission data curve of each monitoring point into several curve segments.

[0050] In the daily operation of chemical plants, precise metering systems and supply processes are in place, and the amount of fossil fuels burned daily is constant. Raw materials such as oil and natural gas are delivered to the combustion equipment at predetermined times, maintaining the stability of the combustion process. Therefore, ideally, if the combustion process is stable and the raw material supply is constant, the combustible gas emission curve should exhibit stable fluctuations with minimal volatility. This stability is a crucial indicator for assessing combustion efficiency and equipment operating status. When relatively large fluctuations occur after a period of stability, it may be due to external factors or anomalies in the combustion process. Further analysis of the specific manifestations of the fluctuations is needed to identify the possible causes and address them to ensure normal operation. Therefore, based on the characteristics of the combustible gas emission curve, the curve is first divided into several segments, each potentially representing a specific combustion state or operating condition. For example, a sudden increase in emissions may indicate decreased combustion efficiency or an abnormal increase in raw material supply; conversely, a decrease in emissions may indicate increased combustion efficiency or a decrease in raw material supply.

[0051] The combustible gas emission data curve segment with a reference duration adjacent to the left of each data point is preset and recorded as the reference range for each data point. In this embodiment, the reference duration is described using 10 minutes as an example.

[0052] The first data point of the combustible gas emission data curve for each monitoring point is recorded as the first segment point of the combustible gas emission data curve for each monitoring point.

[0053] Based on the combustible gas emission data curves of all monitoring points and the reference range for each data point, the segmented evaluation index for each data point of the combustible gas emission data curve of each monitoring point is obtained. The specific calculation formula is as follows:

[0054]

[0055] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point Segmented evaluation indicators for each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point The maximum value of combustible gas emissions between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point The minimum combustible gas emission of each data point between its previous segment point and all other data points. Indicates the first The first curve of combustible gas emission data at each monitoring point The number of all data points between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point Combustible gas emissions at each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point Among all data points between the nth data point and its previous segmentation point, the nth data point is... Combustible gas emissions at each data point; This indicates the total number of monitoring points; Indicates except the first All other monitoring points besides the first monitoring point were at the [number]th monitoring point. The number of monitoring points with segmented points within the reference range of each data point; It is an absolute value function.

[0056] It should be noted that when the first The first curve of combustible gas emission data at each monitoring point The difference between the maximum and minimum values ​​of combustible gas emissions of each data point and the previous segment point. The larger the value, the higher the segmented evaluation index for that data point; Indicates the first The first curve of combustible gas emission data at each monitoring point The mean of the differences in combustible gas emissions between a data point and its previous segment point is the value of the data point's combustible gas emissions. A larger value indicates a higher level of combustible gas emission. The greater the fluctuation range of the curve between a data point and its previous segmentation point, the more unstable it is, and the higher the segmentation evaluation index for that data point; when the... The first curve of combustible gas emission data at each monitoring point There are also segmentation points among other monitoring points near the time of a data point, and the more monitoring points with segmentation points, the higher the segmentation evaluation index of that data point.

[0057] The preset judgment threshold in this embodiment is 0.7. This is used as an example for description. Other values ​​can be set in other embodiments. This embodiment does not limit them.

[0058] when When the value exceeds the preset judgment threshold, the first... The first curve of combustible gas emission data at each monitoring point The data point is used as the first The data curves of combustible gas emissions from each monitoring point are segmented. This process yields several segmentation points for the combustible gas emission data curves of all monitoring points, further dividing the data curve of each monitoring point into several curve segments. It should be noted that the segmentation points for each monitoring point are calculated starting from the first segmentation point, and segmentation evaluation indicators are calculated and determined for each data point. The combustible gas emission data curves of multiple monitoring points are simultaneously processed, with segmentation evaluation indicators calculated and determined for each data point. Specifically, if the reference range for a certain data point exceeds the combustible gas emission data curve, the monitoring time corresponding to the actual existing combustible gas emission data constitutes its reference range.

[0059] Step S003: Based on the standard deviation of combustible gas emission data for all data points in each curve segment of the combustible gas emission data curve at each monitoring point, and the DTW distance between the curve segment and the combustible gas emission data, obtain the standard curve segment of the combustible gas emission data curve at each monitoring point; based on the difference between each curve segment of the combustible gas emission data curve at each monitoring point and its standard curve segment, obtain the degree of data anomaly of each curve segment of the combustible gas emission data curve at each monitoring point.

[0060] In chemical plants, sensors at different monitoring points can be affected by wind, potentially leading to discrepancies in combustible gas emission data at the same monitoring time. Under normal emission conditions, combustible gas emissions should follow a long-term, continuous emission pattern. Significant fluctuations or other changes may indicate actual abnormalities in fuel combustion, or problems caused by environmental factors or monitoring equipment. Therefore, to detect actual anomalies, the first step is to identify a relatively stable curve segment among all segments of the combustible gas emission data curve at each monitoring point, similar to the overall trend of the curve. This segment serves as a standard curve segment, often representing a relatively stable emission state. It is assumed that the combustible gas emission process in the standard curve segment is least affected by external factors. By comparing the differences between the remaining curve segments and the standard curve segment, the interference of external factors on the combustible gas emissions in non-standard curve segments can be quantified, indicating data anomalies. This allows for the differentiation between actual anomalies and false anomalies caused by external factors that require correction.

[0061] Based on the standard deviation of combustible gas emission data for all data points in each curve segment of the combustible gas emission data curve for each monitoring point, and the DTW distance between the curve segment and the combustible gas emission data, the standard evaluation of each curve segment of the combustible gas emission data curve for each monitoring point is obtained. The specific calculation formula is as follows:

[0062]

[0063] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point Standard evaluation of each curve segment; Indicates the first The first curve of combustible gas emission data at each monitoring point The standard deviation of combustible gas emission data for all data points in each curve segment; Indicates the first The combustible gas emission data curve of the first monitoring point and its corresponding first monitoring point The DTW distance of each curve segment.

[0064] It should be noted that the first The first curve of combustible gas emission data at each monitoring point Standard deviation of combustible gas emission data for all data points of each curve segment The value represents the degree of fluctuation; the smaller the value, the higher the stability of the curve segment. Indicates the first The combustible gas emission data curve of the monitoring point and its first The larger the correlation value of the curve segment, the better. The more similar the trend of a curve segment is to the overall curve.

[0065] Among all the curve segments of the combustible gas emission data curve at any monitoring point, the curve segment with the largest standard evaluation is taken as the standard curve segment of the combustible gas emission data curve at that monitoring point.

[0066] On the combustible gas emission data curve of the same monitoring point, the greater the difference between the curve segment and the standard curve segment, the more abnormal the combustible gas emission data of that curve segment is.

[0067] Based on the difference between each segment of the combustible gas emission data curve at each monitoring point and its standard segment, the degree of data anomaly for each segment of the combustible gas emission data curve at each monitoring point is obtained. The specific calculation formula is as follows:

[0068]

[0069] In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point The degree of data anomaly in each curve segment; Indicates the first The first curve of combustible gas emission data at each monitoring point The mean of all combustible gas emission data for each curve segment; Indicates the first The mean of all combustible gas emission data in the standard curve segment of the combustible gas emission data curve for each monitoring point; Indicates the first Standard evaluation of the standard curve segment of the standard curve segment of the combustible gas emission data curve for each monitoring point; Indicates the first The first curve of combustible gas emission data at each monitoring point Standard evaluation of each curve segment; It is an absolute value function.

[0070] It should be noted that when the first The greater the overall difference in combustible gas emissions between a given curve segment and the standard curve segment, the greater the degree of data anomaly in that curve segment; and the... The greater the difference between the standard evaluation of a curve segment and the standard curve segment, the more likely it is to be a data anomaly.

[0071] Step S004: Based on the overlap of all curve segments and their time distributions of the combustible gas emission data curves at all monitoring points, obtain several curve segment groups for the combustible gas emission data curves at all monitoring points; based on the degree of data anomaly and time length of all curve segments within each curve segment group, obtain the wind probability of each curve segment group; based on the wind probability of each curve segment group and the average degree of data anomaly of all curve segments within each curve segment group, obtain all actual abnormal curve segments of the combustible gas emission data curves at all monitoring points.

[0072] When anomalies appear in the data curve, these anomalies may not be entirely caused by actual emissions issues, but rather influenced by external environmental factors. For example, during windy weather, data from all monitoring points will be affected. Monitoring points downwind will show more significant data, while upwind monitoring points may show decreased data. This change may be due to the wind blowing combustible gases downwind from the emission source, causing downwind monitoring points to receive more combustible gases. However, this does not necessarily mean that the actual combustible gas emissions exceed the normal range; it simply means that the data has changed due to the influence of the wind.

[0073] The time range corresponding to any curve segment of the combustible gas emission data curve of any monitoring point, and the reference range of the first and last data points of that curve segment, is all the curve segments of the combustible gas emission data curve of each other monitoring point. This is recorded as the reference curve segment of the combustible gas emission data curve of that monitoring point in each other monitoring point.

[0074] For the combustible gas emission data curve segment of the monitoring point, when there are multiple reference curve segments at any other monitoring point, the reference curve segment with the most time axis overlap is selected.

[0075] Calculate the total time overlap length of each curve segment of the combustible gas emission data curve at each monitoring point with all its reference curve segments. Each curve segment and all its reference curve segments constitute a curve segment combination, and each curve segment combination corresponds to a total time overlap length. If curve segments overlap between curve segment combinations, the curve segment combination with the largest total time overlap length is taken as a curve segment group.

[0076] Based on the degree of data anomaly and the duration of time for all curve segments within each curve segment group, the probability of wind in each curve segment group is obtained. The specific calculation formula is as follows:

[0077]

[0078] In the formula, Indicates the first The probability of wind in each curve segment group; Indicates the first The sum of the differences in time length between any two curve segments in a group of curve segments; Indicates the first The sum of the differences in the degree of data anomaly between any two curve segments in a group of curve segments; This embodiment uses an exponential function with the natural constant as its base. To present the inverse proportional relationship and normalization processing, implementers can set the inverse proportional function and normalization function according to the actual situation.

[0079] It should be noted that the closer the time lengths of all curve segments within a certain curve segment group are, and the closer the corresponding data anomalies are, the greater the likelihood of wind in that curve segment group.

[0080] This embodiment presets a judgment threshold and an anomaly threshold. The judgment threshold is described using 0.7 as an example, and the anomaly threshold is described using 0.6 as an example. When... When the threshold value is greater than or equal to the preset threshold value, the first... The curve segments are caused by wind, which are the normal curve segments; when Less than the preset judgment threshold and the first When the average abnormality of all curve segments in a curve segment group is greater than the preset abnormality threshold, all curve segments in the current curve segment group are considered to be actual abnormal curve segments. Thus, all actual abnormal curve segments of combustible gas emission data curves for all monitoring points are obtained. The greater the abnormality of the curve segment, the more abnormal the current curve segment is.

[0081] In environmental monitoring, when encountering curve segments caused by natural factors such as wind, a rigorous and scientific approach is adopted. For actual anomalies, a rapid response is required, organizing a professional team to conduct on-site investigation and remediation to ensure timely and effective control of emission sources. For false anomalies, monitoring equipment is calibrated and optimized according to the specific circumstances, while data processing methods are adjusted to reduce interference from external factors and improve the accuracy and reliability of the data.

[0082] This concludes the embodiment.

[0083] The present invention also provides a combustible gas monitoring data anomaly detection system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned combustible gas monitoring data anomaly detection method.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A combustible gas monitoring data anomaly detection method characterized by, The method comprises the following steps: Obtain combustible gas emission data curves of each monitoring point; According to the combustible gas emission data curves of all monitoring points and the reference range of each data point, the combustible gas emission data curve of each monitoring point is divided into several curve segments; According to the standard deviation of the combustible gas emission data of all data points in each curve segment of the combustible gas emission data curve of each monitoring point and the DTW distance between the curve segment and the combustible gas emission data, the standard curve segment of the combustible gas emission data curve of each monitoring point is obtained; according to the difference between each curve segment of the combustible gas emission data curve of each monitoring point and the standard curve segment thereof, the data abnormality degree of each curve segment of the combustible gas emission data curve of each monitoring point is obtained; According to the coincidence of all curve segments of the combustible gas emission data curves of all monitoring points and the time distribution thereof, several curve segment groups of the combustible gas emission data curves of all monitoring points are obtained; according to the data abnormality degree and the time length of all curve segments in each curve segment group, the wind blowing possibility of each curve segment group is obtained; according to the wind blowing possibility of each curve segment group and the average data abnormality degree of all curve segments in each curve segment group, all actual abnormal curve segments of the combustible gas emission data curves of all monitoring points are obtained.

2. The combustible gas monitoring data anomaly detection method of claim 1, wherein, The combustible gas emission data curve of each monitoring point is divided into several curve segments according to the combustible gas emission data curves of all monitoring points and the reference range of each data point, and the specific steps include the following: The combustible gas emission data curve segment of the left adjacent reference time length of each data point is preset as the reference range of each data point; The first data point of the combustible gas emission data curve of each monitoring point is recorded as the first segment point of the combustible gas emission data curve of each monitoring point; According to the combustible gas emission data curves of all monitoring points and the reference range of each data point, the segment evaluation index of each data point of the combustible gas emission data curve of each monitoring point is obtained; When the segment evaluation index of any data point of the combustible gas emission data curve of any monitoring point is greater than the preset judgment threshold, the data point of the combustible gas emission data curve of the monitoring point is taken as the segment point of the combustible gas emission data curve of the monitoring point; According to the several segment points of the combustible gas emission data curves of all monitoring points, the combustible gas emission data curve of each monitoring point is divided into several curve segments.

3. The combustible gas monitoring data anomaly detection method of claim 2, wherein, The specific calculation formula of the segment evaluation index of each data point of the combustible gas emission data curve of each monitoring point obtained according to the combustible gas emission data curves of all monitoring points and the reference range of each data point is: In the formula, Indicates the first The first curve of combustible gas emission data at each monitoring point Segmented evaluation indicators for each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point The maximum value of combustible gas emissions between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point The minimum combustible gas emission of each data point between its previous segment point and all other data points. Indicates the first The first curve of combustible gas emission data at each monitoring point The number of all data points between each data point and its previous segment point; Indicates the first The first curve of combustible gas emission data at each monitoring point Combustible gas emissions at each data point; Indicates the first The first curve of combustible gas emission data at each monitoring point Among all data points between the nth data point and its previous segmentation point, the nth data point is... Combustible gas emissions at each data point; This indicates the total number of monitoring points; Indicates except the first All other monitoring points besides the first monitoring point were at the [number]th monitoring point. The number of monitoring points with segmented points within the reference range of each data point; It is an absolute value function.

4. The combustible gas monitoring data anomaly detection method of claim 1, wherein, The specific steps of obtaining the standard curve segment of the combustible gas emission data curve of each monitoring point according to the standard deviation of the combustible gas emission data of all data points in each curve segment of the combustible gas emission data curve of each monitoring point and the DTW distance between the curve segment and the combustible gas emission data include the following: a standard evaluation of each curve segment of the combustible gas emission data curve of each monitoring point is obtained according to the standard deviation of the combustible gas emission data of all data points of each curve segment of the combustible gas emission data curve of each monitoring point and the DTW distance of the curve segment and the combustible gas emission data; in all curve segments of the combustible gas emission data curve of any one monitoring point, the curve segment with the maximum standard evaluation is taken as the standard curve segment of the combustible gas emission data curve of the monitoring point.

5. The combustible gas monitoring data anomaly detection method of claim 4, wherein, the specific calculation formula of the standard evaluation of each curve segment of the combustible gas emission data curve of each monitoring point is as follows: In the formula, represents the standard evaluation of the first curve segment of the combustible gas emission data curve of the first monitoring point; represents the standard evaluation of the first curve segment of the combustible gas emission data curve of the first monitoring point; represents the standard deviation of the combustible gas emission data of all data points of the first curve segment of the combustible gas emission data curve of the first monitoring point; 6. The combustible gas monitoring data anomaly detection method of claim 1, wherein, the specific calculation formula of the data abnormality degree of each curve segment of the combustible gas emission data curve of each monitoring point according to the difference between each curve segment of the combustible gas emission data curve of each monitoring point and the standard curve segment thereof is as follows: In the formula, represents the data anomaly degree of the first curve segment of the combustible gas emission data curve of the first monitoring point; represents the data anomaly degree of the first curve segment of the combustible gas emission data curve of the first monitoring point; represents the mean value of all combustible gas emission data of the first curve segment of the combustible gas emission data curve of the first monitoring point; represents the standard evaluation of the standard curve segment of the combustible gas emission data curve of the first monitoring point; represents the standard evaluation of the first curve segment of the combustible gas emission data curve of the first monitoring point; is an absolute value function.

7. The combustible gas monitoring data anomaly detection method of claim 1, wherein, the specific steps of obtaining a plurality of curve segment groups of the combustible gas emission data curves of all monitoring points according to the coincidence of all curve segments of the combustible gas emission data curves of all monitoring points and the time distribution thereof are as follows: the time range corresponding to the reference range of the first and last data points of any one curve segment of the combustible gas emission data curve of any one monitoring point is recorded as the reference curve segment of the curve segment of the combustible gas emission data curve of the monitoring point in all curve segments of the combustible gas emission data curves of the remaining monitoring points; for the curve segment of the combustible gas emission data curve of the monitoring point, when the reference curve segment of the curve segment in the remaining any one monitoring point is a plurality of curve segments, the reference curve segment with the most time axis coincidence is selected; the total length of time coincidence of each curve segment of the combustible gas emission data curve of each monitoring point and all reference curve segments thereof is calculated, each curve segment and all reference curve segments thereof form a curve segment combination, and each curve segment combination corresponds to a total length of time coincidence; if overlapping curve segments appear between curve segment combinations, the curve segment combination with the maximum total length of time coincidence is taken as a curve segment group.

8. The combustible gas monitoring data anomaly detection method of claim 1, wherein, the specific calculation formula of the wind blowing possibility of each curve segment group according to the data abnormality degree and the time length of all curve segments in each curve segment group is as follows: wherein represents the wind probability of the th curve segment group; represents the sum of the difference values of the time lengths of all arbitrary two curve segments in the th curve segment group; represents the sum of the difference values of the data abnormality degrees of all arbitrary two curve segments in the th curve segment group; is an exponential function with a natural constant as base.

9. The combustible gas monitoring data anomaly detection method of claim 1, wherein, the specific steps of obtaining all actual abnormal curve segments of the combustible gas emission data curves of all monitoring points according to the wind blowing possibility of each curve segment group and the average of the data abnormality degree of all curve segments in each curve segment group are as follows: when the wind blowing possibility of any one curve segment group is less than a preset judgment threshold value and the average of the data abnormality degree of all curve segments of the curve segment group is greater than a preset abnormality threshold value, all curve segments of the curve segment group are actual abnormal curve segments.

10. A combustible gas monitoring data anomaly detection system comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, the computer program is executed by the processor to realize the steps of the combustible gas monitoring data anomaly detection method according to any one of claims 1-9.

Images