Plant carbon flux monitoring data anomaly self-adaptive perception method and system

Abstract

The application discloses a kind of plant carbon flux monitoring data abnormal self-adaptive sensing method and system, specifically related to ecological monitoring data intelligent processing technical field, by collecting frame serial number, receiving time stamp, original sampling time stamp, carbon dioxide mole fraction, water vapor mole fraction and vertical wind speed, establish receiving account book and sampling account book, determine receiving difference and sampling difference, further identify the sampling section to be detected, judge deviation source, write in synchronous boundary, determine closed serial number interval, and complete abnormal segment typing and index output.The application can distinguish receiving link deviation, sampling link deviation and multi-channel synchronization misplacement from the level of time sequence relationship, accurately determine the start and end boundary of abnormal segment, and complete abnormal typing according to peak-valley pairing result and backswing characteristics, form structured index output result, so as to improve the accuracy, stability and subsequent data correction, quality labeling and layered processing of plant carbon flux monitoring data abnormal identification.

Description

Technical Field

[0001] This invention relates to the field of intelligent processing of ecological monitoring data, and in particular to an adaptive sensing method and system for anomalies in plant carbon flux monitoring data. Background Technology

[0002] Plant carbon flux monitoring involves continuous observation and information processing of carbon dioxide exchange processes between ecosystems and the atmosphere. Relevant data typically originates from eddy covariance observations and meteorological and environmental observations related to carbon exchange processes, primarily including multi-source time-series records of carbon dioxide concentration, three-dimensional wind speed, air temperature and humidity, radiation, precipitation, and soil temperature and humidity. The processing of this technology generally includes receiving raw observation data, time synchronization, data organization, quality control, anomaly identification, time period segmentation, correlation analysis, and result output. Its key technical focus is on computer information processing of continuous monitoring data to classify, determine, and manage carbon flux monitoring records.

[0003] In existing technologies, the processing of plant carbon flux monitoring data typically involves first uniformly sampling and aligning the observation records from different sensors. Then, by combining the value range, rate of change, statistical characteristics of the sliding time window, historical seasonal sample distribution, and the correspondence between carbon flux and meteorological factors, abrupt changes, drift values, persistent flat periods, missing segments, and inconsistent records with multiple variables are identified and judged. Existing technologies also establish corresponding data discrimination criteria based on diurnal time periods, seasonal conditions, or station observation rules, and utilize single-class sample reference sets, outlier judgment rules, and continuous sequence comparison methods to perform anomaly detection and classification processing on plant carbon flux monitoring data.

[0004] Current technologies for handling anomalies in monitoring data primarily focus on numerical mutations, missing data compensation, or outlier removal. They lack sufficient differentiation between receiver link deviations, sampling link deviations, and multi-channel synchronization misalignments. Furthermore, they lack a continuous processing mechanism that addresses the boundary determination of anomalous segments, the closure judgment of peak-valley correspondences, and the classification of internal variation patterns within anomalous segments, all centered around temporal relationships. Consequently, this hinders the accurate location and extent of anomalies and makes it difficult to provide a stable basis for subsequent data correction, segment masking, quality labeling, or hierarchical processing.

[0005] Therefore, it is necessary to propose an anomaly adaptive sensing scheme for plant carbon flux monitoring data. Summary of the Invention

[0006] The main objective of this invention is to provide an adaptive sensing method and system for anomalies in plant carbon flux monitoring data. This method uses the carbon dioxide analyzer frame number, three-dimensional ultrasonic anemometer frame number, receiving timestamp, original sampling timestamp, carbon dioxide mole fraction, water vapor mole fraction, and vertical wind speed, which are acquired synchronously during plant carbon flux monitoring, as basic data. Through continuous processing such as ledger establishment, difference identification, boundary writing, peak-valley pairing, closure determination, anomaly classification, and index output, it automatically senses and structures the abnormal segments.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An adaptive sensing method for anomalies in plant carbon flux monitoring data includes: S1. Collect the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed; establish the receiving ledger and the sampling ledger; and determine the receiving difference and the sampling difference. S2. Determine the sampling segment to be inspected and the source of deviation based on the adjacent step size of the received difference, the sampling difference, the carbon dioxide analyzer frame number, and the three-dimensional ultrasonic anemometer frame number; S3. Determine the direction of the carbon dioxide mole fraction and vertical wind speed in the sampling section to be inspected, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, accumulate the sample difference displacement when the received difference and the sampling difference are abnormal at the same time, and write the synchronization boundary when the preset sampling difference threshold is reached. S4. Within the synchronization boundary, the peak and valley points of water vapor mole fraction are paired with the peak and valley points of vertical wind speed, and the closed sequence interval is determined based on the peak point sequence difference, valley point sequence difference and preset sampling difference threshold. S5. Determine the abnormal segment type based on the number of swings of the peak point number difference and valley point number difference within the closed sequence interval and the order of changes in the pairing distance, and output the abnormal segment index.

[0008] Preferably, S1 includes: S11. Collect the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed. Write the corresponding data into a single record according to the same receiving order and assign an incrementing record number to obtain the channel sequence value. S12. Based on the channel sequence value, write the record number, the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, and the receiving timestamp into the receiving ledger; write the record number, the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, and the original sampling timestamp into the sampling ledger; and calculate the difference between adjacent receiving timestamps and the difference between adjacent original sampling timestamps to obtain the ledger time difference value. S13. Based on the ledger time difference value, the difference between adjacent receiving timestamps is taken as the receiving difference, and the difference between adjacent original sampling timestamps is taken as the sampling difference.

[0009] Preferably, S2 includes: S21. Determine the abnormal location and obtain the step size deviation based on the received difference, the sampling difference, the adjacent step sizes of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number. S22. Based on the step size deviation, connect adjacent abnormal positions, superimpose the same-sign region of the received difference and the sampled difference, extract the record number range corresponding to the abnormal position, and remove the position immediately adjacent to the corrected position after a single-point abnormality to obtain the candidate sequence number interval. S23. Based on the interval of the waiting number, count the number of abnormal reception difference, the number of abnormal sampling difference, and the step size jump position of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number within the interval to determine the sampling segment to be inspected and the source of the deviation.

[0010] Preferably, determining the source of deviation includes: When the step size jump occurs in the frame number of the carbon dioxide analyzer or the frame number of the three-dimensional ultrasonic anemometer, and the received difference is abnormal while the sampling difference does not show a corresponding abnormality, the corresponding sampling segment to be inspected is determined as an access link deviation. When the frame number of the carbon dioxide analyzer and the frame number of the three-dimensional ultrasonic anemometer maintain a continuous step size, and the received difference does not show any abnormality while the sampling difference shows continuous abnormality, the corresponding sampling segment to be inspected is determined as a sampling link deviation.

[0011] Preferably, S3 includes: S31. Read the difference between the carbon dioxide mole fraction and the vertical wind speed at adjacent recording positions within the sampling segment to be inspected, determine the number of times in the same direction and the number of times in opposite directions, and obtain the number of directional gauges; S32. Based on the number of direction meters and the total number of records in the sampled segment to be inspected, determine the synchronous offset direction and offset ratio, and convert the offset ratio into integer sample bits to obtain the synchronous offset value; S33. Based on the synchronization offset value, extract the abnormal positions of the received difference and the sampling difference within the sample segment to be inspected, and take the intersection of the recorded positions to form a common anomaly sequence. Accumulate the sample difference displacement along the common anomaly sequence according to the synchronization offset value. Write the synchronization boundary start point when the accumulated value first reaches the preset sampling difference threshold, and write the synchronization boundary end point when the accumulated value falls back to within the preset sampling difference threshold to obtain the synchronization boundary.

[0012] Preferably, the number of direction meters consists of the number of times the direction is in the same direction and the number of times the direction is reversed within the sampling segment to be inspected. The synchronization offset value is a signed integer sample bit calculated based on the proportion of the difference between the number of times the direction is in the same direction and the number of times the direction is reversed relative to the total number of records in the sampling segment to be inspected. When the difference is positive, the synchronization offset value moves forward; when the difference is negative, the synchronization offset value moves backward; and when the difference is zero, the synchronization offset value has zero displacement.

[0013] Preferably, S4 includes: S41. Within the synchronization boundary, according to the increasing order of the record number, each unpaired water vapor mole fraction peak point is paired with the first subsequent vertical wind speed peak point, and each unpaired water vapor mole fraction valley point is paired with the first subsequent vertical wind speed valley point to obtain the peak and valley corresponding sequence number values. S42. Based on the corresponding peak and valley number values, determine the peak number difference between the peak point of the water vapor mole fraction peak point and the peak point of the vertical wind speed, and the valley number difference between the valley point of the water vapor mole fraction valley point and the valley point of the vertical wind speed, and connect the consecutive peak number differences and valley number differences with the same sign to obtain the difference segment value. S43. Based on the difference segment value, the peak point sequence number difference, the valley point sequence number difference, and the preset sampling difference threshold, determine the starting pairing sequence number where both the peak point sequence number difference and the valley point sequence number difference exceed the preset sampling difference threshold, and the ending pairing sequence number where both return to within the preset sampling difference threshold. Then, map the starting pairing sequence number and the ending pairing sequence number to the record sequence number position within the synchronization boundary to obtain the closed sequence number interval.

[0014] Preferably, the closed sequence number interval is determined as follows: the first paired sequence number in which the difference between the peak sequence number and the difference between the valley sequence number have the same sign and the absolute value is greater than the preset sampling difference threshold is selected as the starting sequence number; the first paired sequence number in which the absolute value of the difference between the peak sequence number and the difference between the valley sequence number returns to within the preset sampling difference threshold is retrieved along the direction of subsequent paired sequence numbers as the ending sequence number; when the peak sequence number difference or the valley sequence number difference has a sign reversal between the starting sequence number and the ending sequence number, the corresponding paired sequence number is recorded as the swing position, and the number of times the swing position appears is recorded as the swing number.

[0015] Preferably, S5 includes: S51. Based on the changing order of the peak point number difference, the valley point number difference, the number of oscillations and the spacing between adjacent paired numbers within the closed sequence number interval, determine the sequence code of the paired numbers and obtain the fractal sequence code value. S52. Based on the classification code value, the start number and the end number, determine the segment boundary number and the number of swings to obtain the segment category code; S53. Based on the segment category code, map the segment boundary number to a record number range, and write its start and end record numbers, the segment category code, the number of swings, the interval length, and the sequence code into the index row to obtain the abnormal segment index.

[0016] This invention also discloses an adaptive sensing system for anomaly detection in plant carbon flux monitoring data, comprising: The data acquisition and ledger creation module is used to acquire the frame sequence number of the carbon dioxide analyzer, the frame sequence number of the three-dimensional ultrasonic anemometer, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed, and to create the receiving ledger and the sampling ledger and determine the receiving difference and the sampling difference. The deviation source determination module is used to determine the sampling segment to be inspected and the deviation source based on the adjacent step size of the received difference, the sampling difference, the carbon dioxide analyzer frame number, and the three-dimensional ultrasonic anemometer frame number. The synchronization boundary writing module is used to determine the direction of carbon dioxide mole fraction and vertical wind speed in the sampling segment to be inspected, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, accumulate the sample difference displacement when the received difference and the sampling difference are abnormal at the same time, and write the synchronization boundary when the preset sampling difference threshold is reached. The closed interval determination module is used to pair the peak and valley points of water vapor mole fraction with the peak and valley points of vertical wind speed within the synchronization boundary, and determine the closed number interval based on the peak point number difference, valley point number difference and preset sampling difference threshold. The anomaly classification and index output module is used to determine the anomaly segment classification based on the number of oscillations of the peak point index difference and valley point index difference within the closed index interval, as well as the order of changes in the pairing distance, and output the anomaly segment index. Compared with the prior art, the present invention has the following beneficial effects: 1. This invention introduces both a reception ledger and a sampling ledger to separately model the reception time sequence and the original sampling sequence. It then combines the step size changes of the carbon dioxide analyzer frame sequence number and the 3D ultrasonic anemometer frame sequence number for joint judgment, enabling the differentiation of reception link deviation, sampling link deviation, and synchronization misalignment anomalies resulting from their coupling. Compared to methods that only address numerical mutations, missing measurements, or outliers, this invention identifies the source of anomalies at the temporal relationship level, thereby improving the targeting and accuracy of anomaly identification.

[0017] 2. This invention determines a synchronization offset value within the sampling segment to be inspected, and writes a synchronization boundary when both the received difference and the sampling difference are abnormal. Furthermore, it completes peak and valley point pairing and determines closed sequence intervals within the synchronization boundary, providing a clear basis for determining the start and end ranges of abnormal segments. This not only avoids misjudging scattered fluctuations as continuous anomalies but also avoids fragmenting the same continuous abnormal process, thereby improving the completeness of anomaly localization and the stability of boundary delineation.

[0018] 3. Building upon the identification of abnormal boundaries, this invention further classifies abnormal segments based on the number of oscillations of peak and valley index differences and the order of changes in pairing distances, and outputs an abnormal segment index. This approach moves beyond a simple judgment of whether an anomaly is abnormal; instead, it forms a distinguishable, traceable, and reusable structured representation, facilitating subsequent data correction, segment masking, quality labeling, and hierarchical processing, thus possessing significant engineering application value. Attached Figure Description

[0019] Figure 1 This is an exemplary flowchart of an adaptive sensing method for anomalies in plant carbon flux monitoring data, as shown in some embodiments of the present invention. Figure 2 This is a flowchart of an embodiment of step S1 of the present invention; Figure 3 This is a flowchart of an embodiment of step S2 of the present invention; Figure 4 This is a flowchart of an embodiment of step S3 of the present invention; Figure 5 This is a flowchart of an embodiment of step S4 of the present invention; Figure 6 This is a flowchart of an embodiment of step S5 of the present invention. Detailed Implementation

[0020] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the linguistic context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0021] It should be understood that the terms "system," "device," "unit," and / or "module" as used in this specification are a method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0022] As indicated in this specification and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0023] Flowcharts are used in this specification to illustrate the operations performed by the system according to embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0024] The following is a detailed description of the adaptive sensing method and system for abnormal plant carbon flux monitoring data provided in the embodiments of this specification, with reference to the accompanying drawings.

[0025] Figure 1 This is an exemplary flowchart of an adaptive sensing method for anomalies in plant carbon flux monitoring data, as shown in some embodiments of this specification. In some embodiments, the adaptive sensing method for anomalies in plant carbon flux monitoring data can be executed by processing logic, which may include hardware (e.g., circuitry, dedicated logic, programmable logic, microcode, etc.), software (instructions running on a processing device to execute hardware simulations), and any combination thereof. In some embodiments, Figure 1 One or more operations in the flowchart of the plant carbon flux monitoring data anomaly adaptive sensing method shown can be implemented by a processing device and / or a terminal device. For example, the plant carbon flux monitoring data anomaly adaptive sensing method can be stored in a storage device in the form of a computer program and / or instructions, and invoked and / or executed by the processing device and / or the terminal device.

[0026] like Figure 1 As shown, this invention relates to an adaptive sensing method for anomalies in plant carbon flux monitoring data. This method targets time-series data acquired collaboratively by a carbon dioxide analyzer and a three-dimensional ultrasonic anemometer during plant carbon flux monitoring. It performs joint analysis on the receiving order, original sampling order, frame sequence number progression, and temporal correspondences between multiple physical quantities. The method sequentially completes the establishment of receiving and sampling ledgers, determination of the sampling segments to be inspected, identification of deviation sources, writing of synchronization boundaries, determination of closed sequence number intervals, and classification and indexing of abnormal segments.

[0027] The method specifically includes: S1. Collect the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed; establish the receiving ledger and the sampling ledger; and determine the receiving difference and the sampling difference. S2. Determine the sampling segment to be inspected and the source of deviation based on the adjacent step size of the received difference, the sampling difference, the carbon dioxide analyzer frame number, and the three-dimensional ultrasonic anemometer frame number; S3. Determine the direction of the carbon dioxide mole fraction and vertical wind speed in the sampling section to be inspected, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, accumulate the sample difference displacement when the received difference and the sampling difference are abnormal at the same time, and write the synchronization boundary when the preset sampling difference threshold is reached. S4. Within the synchronization boundary, the peak and valley points of water vapor mole fraction are paired with the peak and valley points of vertical wind speed, and the closed sequence interval is determined based on the peak point sequence difference, valley point sequence difference and preset sampling difference threshold. S5. Determine the abnormal segment type based on the number of swings of the peak point number difference and valley point number difference within the closed sequence interval and the order of changes in the pairing distance, and output the abnormal segment index.

[0028] To ensure that those skilled in the art can implement this method, this embodiment provides a unified description of the input content, judgment conditions, processing order, and output results for each processing step.

[0029] Specifically, such as Figure 2 As shown, in some embodiments of the present invention, step S1 is specifically as follows: During the data acquisition and ledger establishment phase, the system collects the frame sequence number of the carbon dioxide analyzer, the frame sequence number of the 3D ultrasonic anemometer, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed. The corresponding data is written into a single record according to the same receiving order, and an incrementing record number is assigned to each record. Each record retains at least all of the above fields so that subsequent processing can proceed based on the same record sequence. The record number is used to uniformly express the sequential position of each field in the processing chain; the channel sequence value is reflected by this incrementing record number.

[0030] Subsequently, the record number, carbon dioxide analyzer frame number, three-dimensional ultrasonic anemometer frame number, and receiving timestamp are written into the receiving ledger, and the record number, carbon dioxide analyzer frame number, three-dimensional ultrasonic anemometer frame number, and original sampling timestamp are written into the sampling ledger.

[0031] The receiving time interval corresponding to adjacent records in the receiving ledger is used as the receiving difference, and the original sampling time interval corresponding to adjacent records in the sampling ledger is used as the sampling difference. The receiving difference is used to characterize the change in the time position of data when it arrives at the receiving end, and the sampling difference is used to characterize the change in the time position of data on the original sampling side. Whether each difference is abnormal is determined based on the normal sampling cycle of the equipment and in combination with a pre-set allowable fluctuation range; this benchmark and allowable fluctuation range remain consistent within the same processing task.

[0032] The following example further clarifies this step: It should be noted that in this example, the monitoring system continuously acquires 20 records. The system's normal sampling cycle time is set to 50 milliseconds, and the normal frame sequence number step size for both the carbon dioxide analyzer and the 3D ultrasonic anemometer is 1. Anomalies in both the received difference and the sampling difference are determined by deviations from the normal sampling cycle time exceeding the allowable fluctuation range. The preset sampling difference threshold is set to 1 sampling difference unit.

[0033] The sequence code uses a fixed encoding method, where a constant pairing interval is denoted as 0, a decreasing pairing interval as 2, and an increasing pairing interval as 1. The fragment category code uses a fixed encoding method, where unidirectional expansion is denoted as 1, unidirectional reclamation as 2, single swing as 3, and multiple swing as 4.

[0034] In this step, the system collects the frame sequence number of the carbon dioxide analyzer, the frame sequence number of the 3D ultrasonic anemometer, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed, and writes them into a single record according to the same receiving order, assigning an incrementing record number. Subsequently, a receiving ledger and a sampling ledger are established, and the receiving difference and sampling difference are determined respectively.

[0035] The basic ledger data used in this embodiment is shown in Table 1 below: Table 1: Basic Data of the Ledger Based on the above data, the system calculates the time interval sequentially for adjacent records. The reception difference values ​​obtained from the receiving ledger are as follows: 50, 50, 50, 100, 50, 50, 50, 75, 75, 75, 75, 75, 25, 25, 25, 25, 25, 50, 50. The sampling difference values ​​obtained from the sampling ledger are as follows: 50, 50, 50, 50, 50, 50, 50, 75, 75, 75, 75, 75, 25, 25, 25, 25, 25, 50, 50.

[0036] This shows that at record 5, a single-point anomaly occurs in the received difference, while the sampling difference remains normal. At adjacent positions corresponding to records 9 to 18, both the received difference and the sampling difference continuously deviate from the normal sampling rhythm, with records 9 to 13 showing a positive offset and records 14 to 18 showing a callback offset. At this point, the received ledger and sampling ledger are established, and the received difference and sampling difference required for subsequent analysis are obtained.

[0037] like Figure 3 As shown, in some embodiments of the present invention, step S2 specifically includes: During the sampling segment determination phase, the system simultaneously reads the received difference, sampling difference, and the step size changes of the carbon dioxide analyzer frame number and the 3D ultrasonic anemometer frame number between adjacent records. Under normal operating conditions, the frame number should advance continuously according to a preset step size, which can be determined by the device output format, sampling configuration, or debugging results. If the received difference deviates from the normal time interval, the sampling difference deviates from the normal time interval, or any frame number does not advance continuously between adjacent records according to the preset step size, the corresponding record position is recorded as an abnormal position, and a step size deviation is generated. The step size deviation indicates the deviation of the frame number advancement from the normal state; its direction and magnitude are determined with reference to the normal step size.

[0038] For adjacent abnormal locations, the system connects them according to their record numbers. For abnormal locations that are adjacent to each other, with an interval not exceeding the preset connection spacing, and are still in the same abnormal evolution process, they are merged into the same continuous abnormal region. For regions where the received difference and the sampled difference have the same sign, the system overlays them accordingly to highlight the cumulative trend of the two types of difference changes in the same direction.

[0039] Subsequently, the range of record numbers corresponding to the abnormal location is extracted, and the positions immediately following a single-point abnormality that return to normal are removed, resulting in the range of candidate numbers. The position immediately following a single-point abnormality that returns to normal refers to the situation where an abnormality occurs only once, and the subsequent record recovers to the normal judgment range. This return-to-normal position is no longer retained as part of a continuous abnormality.

[0040] It should be noted that during the deviation source determination stage, the system counts the number of abnormal reception difference, the number of abnormal sampling difference, and the step size jump position of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number within the test sequence number interval, and determines the test sampling segment and the deviation source accordingly.

[0041] If the frame number of the carbon dioxide analyzer or the frame number of the three-dimensional ultrasonic anemometer changes by a step size, and the received difference is abnormal while the sampling difference does not show a corresponding abnormality, then the corresponding sampling segment to be tested will be identified as an access link deviation.

[0042] If the frame sequence numbers of the carbon dioxide analyzer and the 3D ultrasonic anemometer maintain a continuous step size, and the received difference is normal while the sampling difference is continuously abnormal, then the corresponding sample segment under inspection is identified as a sampling link deviation. For sample segments under inspection where both the received difference and the sampling difference are abnormal, the subsequent synchronization offset analysis and synchronization boundary writing process will proceed.

[0043] The above-mentioned rules for judging the source of deviation remain unchanged during a complete data processing process to ensure that the same abnormal patterns can be consistently judged.

[0044] For example, in this step, the system determines the abnormal location based on the received difference, the sampling difference, and the adjacent step sizes of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number, connects them to form a candidate inspection number interval, and further determines the sampling segment to be inspected and the source of the deviation.

[0045] In this embodiment, the system first reads the step size of adjacent frame numbers. Comparison shows that between records 4 and 5, the carbon dioxide analyzer frame number jumps from 1004 to 1006, with a step size of 2; within the corresponding time period, the three-dimensional ultrasonic anemometer frame number continues to advance with a step size of 1. At this location, there is also an abnormal received aberration, while the original sampling timestamp continues to advance at the normal pace. Therefore, the system marks the location of record 5 as an abnormal location and obtains that the step size deviation at this point is a positive jump of 1 frame position.

[0046] The system continued searching the corresponding positions of records 9 to 18. It was observed that within this interval, the frame numbers of the carbon dioxide analyzer and the 3D ultrasonic anemometer maintained a continuous step size of 1, with no frame number jumps. However, the received difference and sampling difference were continuously abnormal, exhibiting positive bias with the same sign at positions 9 to 13, and negative bias with the same sign at positions 14 to 18. Based on this, the system connected the corresponding positions of records 9 to 18 into a continuous abnormal region.

[0047] Furthermore, the system organizes the aforementioned abnormal locations. For the single-point anomaly of record 5, since the next record immediately returns to normal, the system, after extracting the record number range corresponding to the anomaly location, removes the immediately adjacent corrected locations, ultimately obtaining the first candidate sequence number interval as the interval containing record 5. For the continuous anomaly region from record 9 to record 18, since the changes in records before and after this interval are continuous, and the received difference and the sampling difference maintain a superposition relationship with the same sign over a relatively long interval, the system retains all records in this segment, obtaining the second candidate sequence number interval as records 9 to 18.

[0048] Subsequently, the system statistically analyzed anomalies within each candidate sequence number interval. For the first candidate sequence number interval, there were jumps in the carbon dioxide analyzer frame sequence number and anomalies in the received difference, while no corresponding anomalies were observed in the sampling difference. Therefore, this segment was identified as an access link deviation. For the second candidate sequence number interval, both the carbon dioxide analyzer frame sequence number and the 3D ultrasonic anemometer frame sequence number maintained continuous step sizes, while both the received difference and sampling difference were continuously abnormal. This indicated that this interval belonged to a sampling segment requiring further synchronization offset analysis. Subsequent steps S3 to S5 all used the aforementioned second candidate sequence number interval, i.e., the data segment corresponding to records 9 to 18, as the processing object.

[0049] like Figure 4 As shown, in some embodiments of the present invention, step S3 is carried out as follows: For the sampled segment to be inspected that enters the synchronous migration analysis, the system reads the changes in carbon dioxide mole fraction and vertical wind speed at adjacent recording positions within the sampled segment, and obtains the previous and subsequent differences for each recording position. The previous difference refers to the change between the current record and the previous record, and the subsequent difference refers to the change between the subsequent record and the current record.

[0050] For cases where the first record only has a subsequent difference and the last record only has a preceding difference, the system uses the difference from one side that is actually obtainable for judgment. For intermediate records, both the preceding and subsequent differences are used for judgment. If the change direction of carbon dioxide mole fraction and vertical wind speed is consistent at the same comparison position, it is recorded as one instance of the same direction; if the change direction is opposite, it is recorded as one instance of the opposite direction; if either change value is zero or insufficient to form a clear direction, then this comparison is not counted in the number of instances of the same direction or the number of instances of the opposite direction. The system accumulates the number of instances of the same direction and the number of instances of the opposite direction within the entire sampling segment to be inspected, and the difference between the two is used as the number of directions.

[0051] Subsequently, the offset percentage is determined based on the proportion of the number of directional gauges relative to the total number of records in the sampled segment to be inspected. This offset percentage is then converted into signed integer sample bits to obtain the synchronization offset value. A fixed rounding rule is used during the conversion, ensuring that the same offset percentage corresponds to a unique integer sample bit under the same processing rule. When the number of directional gauges is positive, the synchronization offset value shifts forward; when the number of directional gauges is negative, the synchronization offset value shifts backward; when the number of directional gauges is zero, the synchronization offset value corresponds to zero displacement.

[0052] It should be noted that the signed integer sample bit referred to here is the offset represented by a record position as the smallest unit. Positive values ​​indicate forward offset, negative values ​​indicate backward offset, and zero values ​​indicate no synchronization offset.

[0053] Furthermore, after obtaining the synchronization offset value, the system extracts the abnormal positions of the received difference and the sampled difference within the sampled segment to be inspected, and forms a common anomaly sequence by taking the intersection of the recorded positions. Along the common anomaly sequence, the system accumulates the sample difference displacement point by point according to the synchronization offset value; when the accumulated value first reaches the preset sample difference threshold, the position is written to the starting point of the synchronization boundary; when the accumulated value falls back to within the preset sample difference threshold in subsequent processing, the corresponding position is written to the ending point of the synchronization boundary, thus obtaining the synchronization boundary.

[0054] It should be noted that the preset sampling difference threshold is used to characterize the degree to which abnormal accumulation has reached a level sufficient to constitute actual synchronization misalignment. Its value can be preset according to the sampling frequency, sensor time resolution, and quality control requirements, and remains unchanged in the same processing task. If the accumulated value does not fall back to within the preset sampling difference threshold before the end of the sampling segment to be inspected, the end position of the sampling segment to be inspected is taken as the synchronization boundary endpoint.

[0055] Once the synchronization boundary is determined, subsequent peak-valley pairing and closure analysis are all performed within this synchronization boundary.

[0056] For example, in this step, the system determines the direction of the carbon dioxide mole fraction and vertical wind speed within the sampled segment to be tested, and determines the synchronization offset value based on the number of times in the same direction and the number of times in the opposite direction; when the received difference and the sampling difference are both abnormal, the sample difference displacement is accumulated, and when it reaches the preset sampling difference threshold, it is written to the synchronization boundary.

[0057] In this embodiment, the values ​​of carbon dioxide mole fraction, water vapor mole fraction, and vertical wind speed recorded from 9 to 20 are shown in Table 2: Table 2: Records 9 to 18 are the sampling segments to be inspected in this step, and records 19 to 20 are used to observe the stabilization status after this segment: Within the sampling segment to be inspected (records 9 to 18), the system compared the preceding and following differences in carbon dioxide mole fraction and vertical wind speed. The comparison results showed that within this sampling segment, the dominant change directions of carbon dioxide mole fraction and vertical wind speed were consistent, exhibiting a synchronous evolution characteristic of "rising-falling-rising again." The number of same-direction judgments significantly exceeded the number of reverse-direction judgments, leading the system to determine that the direction count for this segment was positive, and thus the synchronous offset direction was determined to be forward. Following the fixed conversion rule used in this embodiment, the system converted the offset proportion of this sampling segment to one positive sample position; therefore, the synchronous offset value was determined to be one positive sample position.

[0058] Based on this, the system extracts the abnormal locations of the received difference and the abnormal locations of the sampling difference within the sampled segment to be inspected. As shown in step S1, both received difference and sampling difference anomalies exist simultaneously at positions corresponding to records 9 to 18, thus their intersection constitutes a common anomaly sequence. The system accumulates the sample difference displacement point by point along this common anomaly sequence. At positions corresponding to records 9 to 13, since both the received difference and the sampling difference show positive shifts, the cumulative displacement continuously increases; at positions corresponding to records 14 to 18, since the received difference and the sampling difference enter a callback phase, the cumulative displacement gradually decreases. According to the preset sampling difference threshold set in this embodiment, when the cumulative value first reaches the threshold at record 9, the system writes record 9 to the starting point of the synchronization boundary; subsequently, it continues to track the change of the cumulative value, and when the cumulative value falls back to within the threshold at record 18, the system writes record 18 to the ending point of the synchronization boundary. Therefore, the synchronization boundary obtained in this embodiment is records 9 to 18.

[0059] like Figure 5 As shown, in some embodiments of the present invention, step S4 is carried out as follows: First, within the synchronization boundary, the system identifies peaks and troughs in the water vapor mole fraction and vertical wind speed sequences, respectively. A peak is a local high point where, relative to its preceding and following records, the value changes from increasing to decreasing; a trough is a local low point where, relative to its preceding and following records, the value changes from decreasing to increasing. To avoid the influence of local noise, the system can perform uniform smoothing before peak and trough identification, but the same identification caliber must be used for both water vapor mole fraction and vertical wind speed. For cases where multiple consecutive records have equal values ​​forming a plateau, the starting point of the plateau or a location selected according to a unified rule is used as the corresponding peak or trough, and this consistency is maintained within the same processing task.

[0060] After peak and valley identification is completed, each unpaired water vapor mole fraction peak is paired with the first unpaired vertical wind speed peak following it, and each unpaired water vapor mole fraction valley is paired with the first unpaired vertical wind speed valley following it, in ascending order of record number, to obtain the corresponding peak and valley sequence values. It should be noted that the "first following" here refers to the first vertical wind speed peak or valley that meets the corresponding peak or valley type condition and has not yet participated in the same type of pairing, found in ascending direction of record number, starting from the current water vapor mole fraction peak or valley.

[0061] After completing the peak-valley pairing, the system determines the peak number difference between the peak point of water vapor mole fraction and the peak point of vertical wind speed, and the valley number difference between the valley point of water vapor mole fraction and the valley point of vertical wind speed.

[0062] Specifically, a positive peak or valley index difference indicates that the corresponding vertical wind speed peak or valley is located after the corresponding water vapor mole fraction peak or valley; a negative difference indicates it is located before; and a zero difference indicates that they are at the same recording position or considered to be at the same position. Subsequently, consecutive peak index differences with the same sign are concatenated, and consecutive valley index differences with the same sign are concatenated to obtain the difference segment value. It should be noted that "consecutive with the same sign" here means that the peak or valley index differences within adjacent paired index ranges maintain the same sign, and do not return to within the preset sampling difference threshold or undergo sign reversal within this range. The difference segment value is used to characterize the degree of continuation of peak or valley offset in a certain direction within a consecutive paired range.

[0063] During the process of determining the closed sequence interval, the system makes a joint judgment based on the difference segment value, the peak point sequence difference, the valley point sequence difference, and the preset sampling difference threshold.

[0064] In the specific processing, the first pairing sequence number where the difference between the peak and valley sequence numbers has the same sign and their absolute values ​​are both greater than the preset sampling difference threshold is selected as the starting pairing sequence number; then, the search is carried out along the direction of the subsequent pairing sequence numbers to find the first pairing sequence number where the absolute values ​​of the difference between the peak and valley sequence numbers return to within the preset sampling difference threshold, which is selected as the ending pairing sequence number; then, the starting pairing sequence number and the ending pairing sequence number are mapped to the record sequence number position within the synchronization boundary to obtain the closed sequence number interval.

[0065] Of course, if no matching sequence number returns to within the preset sampling difference threshold during the retrieval process until the end of the synchronization boundary, the last valid matching sequence number within the synchronization boundary will be used as the end matching sequence number.

[0066] In this way, the closed sequence interval begins with both peak and valley shifts crossing the threshold simultaneously and ends with both shifts returning to the threshold simultaneously, thus reflecting a continuous misalignment process more completely.

[0067] It should also be noted that in this embodiment, within the closed sequence interval, the system continues to determine whether the peak point sequence difference and valley point sequence difference have swung back. Specifically, the sign state corresponding to the peak point sequence difference and valley point sequence difference at the beginning of the closed sequence interval is taken as the starting state, and then each subsequent pairing sequence is compared. If the peak point sequence difference or valley point sequence difference at a certain subsequent pairing sequence number reverses its sign relative to the starting state, then the pairing sequence number is recorded as the swing position. If multiple adjacent pairing sequence numbers continuously maintain the same reversed state, then only the first pairing sequence number of the continuous reverse interval is counted as one swing, and the same continuous swing process is not counted repeatedly. The number of times all swing positions appear within the closed sequence interval is the number of swings. Thus, the number of swings represents the number of times the offset direction inside the abnormal segment has been effectively reversed, rather than simply the number of fluctuation points.

[0068] For example, in this step, the system pairs the peak and valley points of water vapor mole fraction with the peak and valley points of vertical wind speed within the synchronization boundary, and determines the closed sequence interval based on the difference in peak point number, the difference in valley point number, and the preset sampling difference threshold.

[0069] In this embodiment, the synchronization boundary is records 9 to 18. Within this boundary, the system extracts the peaks and valleys of the water vapor mole fraction sequence and the vertical wind speed sequence. After extraction, the water vapor mole fraction forms a continuous peak-valley alternation structure within this boundary, and the vertical wind speed also forms a corresponding peak-valley alternation structure, but overall, it lags behind the water vapor mole fraction.

[0070] The system pairs each unpaired peak point of water vapor mole fraction with the first subsequent peak point of vertical wind speed, and pairs each unpaired valley point of water vapor mole fraction with the first subsequent valley point of vertical wind speed, in ascending order of record number. After sequential pairing, four sets of peak-valley correspondence results directly related to the closure judgment are obtained, as shown in Table 3.

[0071] Table 3: Peak-valley correspondence results directly related to closure judgment in 4 groups: As shown in Table 3, at pairing number 1, the peak difference and valley difference have the same sign, and their absolute values ​​are both greater than the preset sampling difference threshold of 1 set in this embodiment. Therefore, the system uses pairing number 1 as the starting pairing number of the closed number interval. Subsequently, the search continues along the direction of subsequent pairing numbers. It can be seen that by pairing number 4, the absolute values ​​of both the peak difference and valley difference have returned to within the threshold. Therefore, the system uses pairing number 4 as the ending pairing number of the closed number interval.

[0072] Furthermore, the system checks the sign changes of the peak and valley number differences between the starting and ending pairing numbers. The check shows that within pairing numbers 1 to 4, all peak and valley number differences remain positive, with no sign reversal. Therefore, there is no swingback position in this embodiment, corresponding to a swingback count of 0. Finally, the system maps the starting and ending pairing numbers back to the record number positions within the synchronization boundary, resulting in a closed sequence number interval corresponding to records 10 to 18 in this embodiment.

[0073] like Figure 6 As shown, in some embodiments of the present invention, step S5 includes: During the anomaly classification phase, the system determines the sequence code of the corresponding pairing number based on the changes in the peak point sequence difference, valley point sequence difference, number of oscillations, and adjacent pairing distances within the closed sequence number interval. The pairing distance reflects the degree of misalignment at each pairing position and can be characterized by both the peak point sequence difference and the valley point sequence difference. The system compares the changes in the pairing distance between adjacent pairing positions: when the pairing distance of the next pairing position increases significantly, it is recorded as an increasing sequence code; when the pairing distance of the next pairing position decreases significantly, it is recorded as a decreasing sequence code; when the pairing distance between two adjacent pairing positions remains basically unchanged, it is recorded as a maintaining sequence code.

[0074] To facilitate subsequent encoding, in this embodiment, the enlarging sequence code is denoted as "1", the decreasing sequence code as "2", and the maintaining sequence code as "0". The sequence codes are arranged sequentially according to their pairing order to form a sequence code sequence.

[0075] Furthermore, during the generation of the fractal code value, the system simultaneously considers the same sign direction of the peak and valley index differences at the initial pairing index, the changing trend represented by the sequence code sequence, and the number of oscillations. To ensure the determinism of the fractal result, in this embodiment, when the initial same sign direction is positive, the direction code "1" is used; when the initial same sign direction is negative, the direction code "2" is used. The sequence code sequence exhibits continuous expansion, trend code "1" is used; when it exhibits continuous contraction, trend code "2" is used; when it exhibits expansion followed by contraction, trend code "3" is used; when it exhibits contraction followed by expansion, trend code "4" is used; and when it exhibits multiple oscillations, trend code "5" is used. When the number of oscillations is zero, the oscillation code "0" is used; when the number of oscillations is one, the oscillation code "1" is used; and when the number of oscillations is two or more, the oscillation code "2" is used. The system obtains the fractal code value by combining the "direction code, trend code, and oscillation code" in a fixed order. Using the above fixed combination method, closed index intervals of the same type obtain consistent fractal code values ​​under the same rules.

[0076] During the fragment category code determination process, the system determines the fragment boundary number and the number of swings based on the classification sequence code value, the starting pairing sequence number, and the ending pairing sequence number, and obtains the fragment category code. To ensure clear and unambiguous classification results, in this embodiment, unidirectional expansion-type abnormal fragments correspond to fragment category code "1", unidirectional recovery-type abnormal fragments correspond to fragment category code "2", single swing-type abnormal fragments correspond to fragment category code "3", and multiple swing-type abnormal fragments correspond to fragment category code "4". When the fragment category code is "1" or "2", the fragment boundary number is determined by the starting pairing sequence number and the ending pairing sequence number; when the fragment category code is "3", the fragment boundary number is determined by the starting pairing sequence number, the first swing position, and the ending pairing sequence number; when the fragment category code is "4", the fragment boundary number is jointly determined by the starting pairing sequence number, all swing positions, and the ending pairing sequence number.

[0077] Therefore, the fragment category code not only reflects the overall type of the abnormal fragment, but also determines how the internal structure of the abnormal fragment is preserved during index output.

[0078] During the index output phase, the system maps the segment boundary number to a record number range based on the segment category code, and writes the start and end record numbers, segment category code, number of swings, interval length, and sequence code of the abnormal segment into the index row to obtain the abnormal segment index. The start and end record numbers are used to locate the position of the abnormal segment in the original monitoring data, the segment category code is used to indicate the type of the abnormal segment, the number of swings is used to indicate the inversion complexity of the internal offset direction of the abnormal segment, the interval length is used to indicate the duration range of the abnormal segment, and the sequence code is used to preserve the chronological order of the misalignment evolution within the abnormal segment.

[0079] It should be noted that when a closed sequence number interval corresponds to multiple segment boundary numbers, the system can generate index rows according to the range between adjacent boundary numbers, or generate index rows containing internal structural information according to the entire closed sequence number interval. However, a unified output standard should be used in a single processing task.

[0080] For example, in this step, the system determines the abnormal segment type based on the number of swings of the peak point index difference and valley point index difference within the closed sequence interval and the order of changes in the pairing distance, and outputs the abnormal segment index.

[0081] In this embodiment, the four sets of peak-valley correspondences within the closed sequence interval are sequentially represented as "2 / 2, 2 / 2, 2 / 1, 1 / 1". The system compares the changes in pairing distance between adjacent pairs accordingly: from group 1 to group 2, the pairing distance remains unchanged, so the sequence code for this step is recorded as 0; from group 2 to group 3, the pairing distance decreases, so the sequence code for this step is recorded as 2; from group 3 to group 4, the pairing distance continues to decrease, so the sequence code for this step is still recorded as 2. Therefore, the sequence code sequence obtained in this embodiment is "0-2-2".

[0082] Meanwhile, since the difference between the peak and valley indexes at the beginning of the closed sequence interval is positive, it indicates that the initial direction of the abnormal segment is positively lagging. Furthermore, since the sequence code sequence as a whole exhibits a unidirectional convergence characteristic from large to small, and the number of oscillations is 0, the system determines this closed sequence interval as a unidirectional recycling-type abnormal segment. According to the preset classification coding rules of this embodiment, the direction code is positively encoded, the trend code is continuously recycled encoded, and the oscillation code is 0, resulting in the classification code value "120". Further, according to the segment category mapping rules, the system determines this segment as segment category code 2.

[0083] After determining the fragment category code, the system maps the start and end positions of the closed sequence number interval to a record sequence number range, and forms an abnormal fragment index. In this embodiment, the abnormal fragment index writing results are shown in Table 4.

[0084] Table 4: Results of writing abnormal fragment indexes: Thus, this embodiment completes the entire processing flow from the original monitoring records, through ledger establishment, determination of sampled segments to be inspected, writing of synchronization boundaries, determination of closed sequence number intervals, to the classification and indexing of abnormal segments. Specifically, the single-point anomaly corresponding to record 5 is identified as an access link deviation; the data segments corresponding to records 9 to 18 are identified as sampled segments to be inspected that require further synchronization analysis, and finally, a one-way recovery-type abnormal segment index is output within the range of records 10 to 18.

[0085] This invention also provides an adaptive sensing system for anomaly detection in plant carbon flux monitoring data, used to execute the above-described method. The system includes a data acquisition and ledger establishment module, a deviation source determination module, a synchronization boundary writing module, a closed interval determination module, and an anomaly classification and index output module.

[0086] The data acquisition and ledger creation module is used to collect the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed, and to create the receiving ledger and the sampling ledger and determine the receiving difference and the sampling difference. The deviation source determination module is used to determine the sampling segment to be inspected and the deviation source based on the received difference, the sampling difference, and the adjacent step size of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number. The synchronization boundary writing module is used to determine the direction of carbon dioxide mole fraction and vertical wind speed in the sampled section to be tested, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, and accumulate the sample difference displacement when the received difference and the sampling difference are both abnormal, and write the synchronization boundary when the preset sampling difference threshold is reached. The closed interval determination module is used to pair the peak and valley points of water vapor mole fraction with the peak and valley points of vertical wind speed within the synchronization boundary, and determine the closed sequence interval based on the difference of peak point number, the difference of valley point number and the preset sampling difference threshold. The anomaly classification and index output module is used to determine the anomaly segment classification based on the number of oscillations of the peak point index difference and valley point index difference within the closed sequence interval and the order of changes in the pairing distance, and output the anomaly segment index.

[0087] The above modules can be implemented by the processor executing the corresponding program, or by the hardware logic unit with the corresponding function.

[0088] It should be noted that the normal sampling beat, frame sequence number preset step size, reception difference anomaly judgment threshold, sampling difference anomaly judgment threshold, preset sampling difference threshold, abnormal position connection spacing, peak and valley identification caliber, synchronization offset value rounding rule, and specific encoding caliber of sequence code, classification sequence code value, and fragment category code involved in this embodiment can all be preset according to the equipment model, sampling frequency, and monitoring task requirements, but should remain consistent during the same batch of data processing. As long as it is still based on the difference judgment between the reception ledger and the sampling ledger, the synchronization offset is determined based on the multi-channel directional relationship, the closed sequence number interval is determined based on peak and valley pairing, and the abnormal fragment classification and index output is completed based on the number of swings and the change order of pairing spacing, all belong to the implementation mode of the technical idea described in this invention.

[0089] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0090] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Moreover, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0091] Furthermore, those skilled in the art will understand that various aspects of this specification can be described and illustrated in several patentable ways, including any new and useful combinations of processes, machines, products, or substances, or any new and useful improvements thereof. Accordingly, various aspects of this specification can be implemented entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. All of the above hardware or software may be referred to as a “data block,” “module,” “engine,” “unit,” “component,” or “system.” Furthermore, various aspects of this specification may be represented as a computer product located on one or more computer-readable media, including computer-readable program code.

[0092] Computer storage media may contain a propagated data signal containing computer program encoding, for example, on baseband or as part of a carrier wave. This propagated signal may take various forms, including electromagnetic, optical, and suitable combinations thereof. Computer storage media can be any computer-readable medium other than a computer-readable storage medium, which can be connected to an instruction execution system, apparatus, or device to enable communication, propagation, or transmission of a program for use. The program encoding located on the computer storage medium can be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or similar media, or any combination of the above media.

[0093] The computer program code required for the operation of each part of this manual can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc.; conventional procedural programming languages ​​such as C, Visual Basic, Fortran2003, Perl, COBOL2002, PHP, ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or processing device. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0094] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments described herein. For example, while the system components described above can be implemented by hardware devices, they can also be implemented solely by software solutions, such as installing the described system on existing processing devices or mobile devices.

[0095] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0096] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. An adaptive sensing method for anomalies in plant carbon flux monitoring data, characterized in that, include: S1. Collect the frame number of the carbon dioxide analyzer, the frame number of the three-dimensional ultrasonic anemometer, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed. Establish the receiving ledger and the sampling ledger and determine the receiving difference and the sampling difference. Write the corresponding data into the record according to the same receiving order and assign an incrementing record number. S2. Based on the adjacent step sizes of the received difference, the sampling difference, the carbon dioxide analyzer frame number, and the three-dimensional ultrasonic anemometer frame number, determine the abnormal location and obtain the step size deviation. Based on the step size deviation, connect adjacent abnormal locations, superimpose the same-sign regions of the received difference and the sampling difference, extract the record number range corresponding to the abnormal location, and remove the positions immediately adjacent to the corrected position after a single-point abnormality to obtain the candidate inspection number interval. Based on the candidate inspection number interval, count the number of abnormal received difference, the number of abnormal sampling difference, and the step size jump positions of the carbon dioxide analyzer frame number and the three-dimensional ultrasonic anemometer frame number within the interval to determine the sampling segment to be inspected and the source of deviation. S3. Determine the direction of the carbon dioxide mole fraction and vertical wind speed in the sampling section to be inspected, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, accumulate the sample difference displacement when the received difference and the sampling difference are abnormal at the same time, and write the synchronization boundary when the preset sampling difference threshold is reached. S4. Within the synchronization boundary, the peak and valley points of water vapor mole fraction are paired with the peak and valley points of vertical wind speed, and the closed sequence interval is determined based on the peak point sequence difference, valley point sequence difference and preset sampling difference threshold. S5. Determine the abnormal segment type based on the number of swings of the peak point number difference and valley point number difference within the closed sequence interval and the order of changes in the pairing distance, and output the abnormal segment index.

2. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 1, characterized in that, S1 includes: S11. Collect the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed. Write the corresponding data into a single record according to the same receiving order and assign an incrementing record number to obtain the channel sequence value. S12. Based on the channel sequence value, write the record number, the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, and the receiving timestamp into the receiving ledger; write the record number, the carbon dioxide analyzer frame number, the three-dimensional ultrasonic anemometer frame number, and the original sampling timestamp into the sampling ledger; and calculate the difference between adjacent receiving timestamps and the difference between adjacent original sampling timestamps to obtain the ledger time difference value. S13. Based on the ledger time difference value, the difference between adjacent receiving timestamps is taken as the receiving difference, and the difference between adjacent original sampling timestamps is taken as the sampling difference.

3. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 1, characterized in that, The determination of the source of deviation includes: When the step size jump occurs in the frame number of the carbon dioxide analyzer or the frame number of the three-dimensional ultrasonic anemometer, and the received difference is abnormal while the sampling difference does not show a corresponding abnormality, the corresponding sampling segment to be inspected is determined as an access link deviation. When the frame number of the carbon dioxide analyzer and the frame number of the three-dimensional ultrasonic anemometer maintain a continuous step size, and the received difference does not show any abnormality while the sampling difference shows continuous abnormality, the corresponding sampling segment to be inspected is determined as a sampling link deviation.

4. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 1, characterized in that, S3 includes: S31. Read the difference between the carbon dioxide mole fraction and the vertical wind speed at adjacent recording positions within the sampling segment to be inspected, determine the number of times in the same direction and the number of times in opposite directions, and obtain the number of directional gauges; S32. Based on the number of direction meters and the total number of records in the sampled segment to be inspected, determine the synchronous offset direction and offset ratio, and convert the offset ratio into integer sample bits to obtain the synchronous offset value; S33. Based on the synchronization offset value, extract the abnormal positions of the received difference and the sampling difference within the sample segment to be inspected, and take the intersection of the recorded positions to form a common anomaly sequence. Accumulate the sample difference displacement along the common anomaly sequence according to the synchronization offset value. Write the synchronization boundary start point when the accumulated value first reaches the preset sampling difference threshold, and write the synchronization boundary end point when the accumulated value falls back to within the preset sampling difference threshold to obtain the synchronization boundary.

5. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 4, characterized in that, The number of directional indicators consists of the number of times they are in the same direction and the number of times they are in the opposite direction within the sample segment to be inspected. The synchronization offset value is a signed integer sample bit calculated based on the proportion of the difference between the number of times they are in the same direction and the number of times they are in the opposite direction relative to the total number of records in the sample segment to be inspected. When the difference is positive, the synchronization offset value moves forward; when the difference is negative, the synchronization offset value moves backward; and when the difference is zero, the synchronization offset value has zero displacement.

6. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 1, characterized in that, S4 includes: S41. Within the synchronization boundary, according to the increasing order of the record number, each unpaired water vapor mole fraction peak point is paired with the first subsequent vertical wind speed peak point, and each unpaired water vapor mole fraction valley point is paired with the first subsequent vertical wind speed valley point to obtain the peak and valley corresponding sequence number values. S42. Based on the corresponding peak and valley number values, determine the peak number difference between the peak point of the water vapor mole fraction peak point and the peak point of the vertical wind speed, and the valley number difference between the valley point of the water vapor mole fraction valley point and the valley point of the vertical wind speed, and connect the consecutive peak number differences and valley number differences with the same sign to obtain the difference segment value. S43. Based on the difference segment value, the peak point sequence number difference, the valley point sequence number difference, and the preset sampling difference threshold, determine the starting pairing sequence number where both the peak point sequence number difference and the valley point sequence number difference exceed the preset sampling difference threshold, and the ending pairing sequence number where both return to within the preset sampling difference threshold. Then, map the starting pairing sequence number and the ending pairing sequence number to the record sequence number position within the synchronization boundary to obtain the closed sequence number interval.

7. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 6, characterized in that, The closed sequence number interval is determined as follows: the first paired sequence number in which the difference between the peak sequence number and the difference between the valley sequence number have the same sign and the absolute value is greater than the preset sampling difference threshold is selected as the starting sequence number; the first paired sequence number in which the absolute value of the difference between the peak sequence number and the difference between the valley sequence number returns to within the preset sampling difference threshold is retrieved along the direction of subsequent paired sequence numbers as the ending sequence number; when the peak sequence number difference or the valley sequence number difference has a sign reversal between the starting sequence number and the ending sequence number, the corresponding paired sequence number is recorded as the swing position, and the number of times the swing position appears is recorded as the swing number.

8. The adaptive sensing method for anomalies in plant carbon flux monitoring data according to claim 7, characterized in that, S5 includes: S51. Based on the changing order of the peak point number difference, the valley point number difference, the number of oscillations and the spacing between adjacent paired numbers within the closed sequence number interval, determine the sequence code of the paired numbers and obtain the fractal sequence code value. S52. Based on the classification code value, the start number and the end number, determine the segment boundary number and the number of swings to obtain the segment category code; S53. Based on the segment category code, map the segment boundary number to a record number range, and write its start and end record numbers, the segment category code, the number of swings, the interval length, and the sequence code into the index row to obtain the abnormal segment index.

9. An adaptive sensing system for anomaly detection in plant carbon flux monitoring data, used to execute the method according to any one of claims 1-8, characterized in that, include: The data acquisition and ledger creation module is used to acquire the frame sequence number of the carbon dioxide analyzer, the frame sequence number of the three-dimensional ultrasonic anemometer, the receiving timestamp, the original sampling timestamp, the carbon dioxide mole fraction, the water vapor mole fraction, and the vertical wind speed, and to create the receiving ledger and the sampling ledger and determine the receiving difference and the sampling difference. The deviation source determination module is used to determine the sampling segment to be inspected and the deviation source based on the adjacent step size of the received difference, the sampling difference, the carbon dioxide analyzer frame number, and the three-dimensional ultrasonic anemometer frame number. The synchronization boundary writing module is used to determine the direction of carbon dioxide mole fraction and vertical wind speed in the sampling segment to be inspected, determine the synchronization offset value according to the number of times in the same direction and the number of times in the opposite direction, accumulate the sample difference displacement when the received difference and the sampling difference are abnormal at the same time, and write the synchronization boundary when the preset sampling difference threshold is reached. The closed interval determination module is used to pair the peak and valley points of water vapor mole fraction with the peak and valley points of vertical wind speed within the synchronization boundary, and determine the closed number interval based on the peak point number difference, valley point number difference and preset sampling difference threshold. The anomaly classification and index output module is used to determine the anomaly segment classification based on the number of oscillations of the peak point index difference and the valley point index difference within the closed sequence interval, as well as the order of changes in the pairing distance, and output the anomaly segment index.

Images