A durian sprinkling irrigation system fault monitoring method based on anomaly detection

Abstract

The application discloses a durian sprinkling irrigation system fault monitoring method based on abnormality detection, relates to the technical field of intelligent irrigation detection, and comprises the following steps: collecting sprinkling irrigation operation data, uniformly marking the sprinkling irrigation operation data with time and source, and generating an operation access sequence; tracking a deviation characteristic sequence in time, obtaining a control deviation transmission path, analyzing a deviation transmission sequence and an expansion trend through a time sequence evolution tracking method, determining a control imbalance starting position and a starting direction, and generating an abnormality propagation sequence; performing reverse positioning analysis on the abnormality propagation sequence, obtaining an abnormality source position, combining the abnormality source position with a state change characteristic, determining a fault source and an influence range, and generating a fault determination scheme. The application improves abnormality recognition, propagation analysis and fault positioning capabilities.

Description

Technical Field

[0001] This invention relates to the field of intelligent irrigation detection technology, and in particular to a fault monitoring method for a durian sprinkler irrigation system based on anomaly detection. Background Technology

[0002] With the development of facility agriculture and orchard precision management technology, durian cultivation is gradually shifting from traditional experience-based irrigation to sprinkler irrigation based on sensor perception and automatic control. Existing sprinkler irrigation methods typically collect key parameters in real time during the irrigation process by deploying flow sensors, pressure sensors, and soil moisture sensors, and combine them with control strategies to achieve timed or on-demand water supply control. On this basis, some methods further introduce data monitoring and fault identification mechanisms, which identify problems such as sprinkler head blockage, pipeline leakage, or abnormal pressure by performing threshold judgment, trend analysis, or simple model comparison on the collected operating data.

[0003] When analyzing the operation status of sprinkler irrigation, the focus is often on the trend analysis of a single parameter within a certain time segment. Although this can reflect abnormal situations at a certain moment or in a certain dimension, it is still insufficient in terms of the coupling relationship of multiple parameters and their continuous evolution over time. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a fault monitoring method for durian sprinkler irrigation systems based on anomaly detection, which solves the problems of difficulty in uniformly characterizing the multi-parameter coupling relationship and difficulty in continuously tracking the anomaly evolution process.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a fault monitoring method for a durian sprinkler irrigation system based on anomaly detection, comprising: collecting sprinkler operation data and unifying the time and source identifiers of the sprinkler operation data to generate an operation access sequence; constructing a time correlation of the operation access sequence to obtain water operation parameters, and performing time location matching and consistency analysis on the water operation parameters to generate an operation consistency sequence; comparing adjacent time location changes in the operation consistency sequence to obtain change correlation features, identifying deviations between flow rate information, pressure information, and soil moisture information from the change correlation features, and using a coupled deviation mining method to perform time-series propagation merging of deviations to generate a deviation feature sequence; performing time-series inheritance tracking on the deviation feature sequence to obtain the control deviation propagation path, and using a time-series evolution tracking method to analyze the deviation propagation order and expansion trend, determine the starting position and starting direction of control imbalance, and generate an anomaly propagation sequence; performing reverse localization analysis on the anomaly propagation sequence to obtain the anomaly source location, combining the anomaly source location with state change features, determining the fault source and impact range, and generating a fault determination scheme.

[0007] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for collecting sprinkler irrigation operation data and unifying the time and source identifiers of the sprinkler irrigation operation data to generate an operation access sequence are as follows: Collect sprinkler irrigation operation data, read the sprinkler irrigation operation data item by item and sort it by time to generate an initial record set; Source identifier data is extracted from the initial record set, the source identifier data is combined and marked to obtain the source identifier record set, and the collection time of the source identifier record set is uniformly calibrated and time-aligned to generate a time-unified record set. The unified time record set is arranged continuously in chronological order to obtain the access-related record set. The access-related record set is then integrated and organized in a unified manner to generate the running access sequence.

[0008] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for constructing a time correlation of the operating access sequence to obtain water operation parameters are as follows: The operational access sequences are grouped and expanded according to the source identifier data to obtain the location number mapping value. The location number mapping value is then merged and filtered to generate a set of candidate location records. Flow rate, pressure and soil moisture information are extracted from the candidate location record set to construct a candidate set of moisture parameters; The temporal order and positional correspondence of adjacent records in the candidate set of moisture parameters are matched and verified to obtain the time-position matching sequence. The time-position matching sequence is then combined and organized to generate moisture operation parameters.

[0009] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for performing time-location matching and consistency analysis of water operation parameters to generate a consistent operation sequence are as follows: Consistency calculations are performed on the water operation parameter group to obtain a consistency parameter set. Based on the consistency parameter set, the water operation parameters are continuously arranged and their continuity is determined to generate a continuous consistency segment set. The continuous consistent segment set is connected, sorted, and anomaly removed to obtain a consistent segment sorting set. The consistent segment sorting set is then uniformly organized to generate a running consistent sequence.

[0010] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for comparing adjacent time position changes of the consistent running sequence to obtain change correlation features are as follows: Expand the points of the consistent running sequence according to time order to generate points at the same location; Construct neighboring time records based on the same location points, extract moisture change coupling features from the neighboring time records, and perform difference comparison and correlation integration on the moisture change coupling features to generate change correlation features.

[0011] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the deviation between flow rate information, pressure information, and soil moisture information is identified from the change correlation features, and a coupled deviation mining method is used to perform time-series recursive merging of the deviations to generate a deviation feature sequence. The specific steps are as follows: By comparing the direction and magnitude of change in the change correlation features item by item, we can identify change items with inconsistent directions, determine the dominant deviation relationship, and generate deviation information. The coupled deviation mining method is used to check the consistency of the same location, the consistency of the deviation relationship, and the change continuity of the deviation along the time direction, and to generate a set of progressive deviation segments. The time continuity and deviation stability of the propagation deviation segment set are checked to obtain the propagation check segment set. The broken segments in the propagation check segment set are segmented and organized to generate the deviation feature sequence.

[0012] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for performing time-series tracing of the deviation feature sequence to obtain the control deviation propagation path are as follows: The deviation feature sequence is expanded segment by segment to obtain the preceding and following connecting segments. The deviation segment feature information is extracted from the preceding and following connecting segments. The connecting strength of the deviation segment feature information is determined to obtain the connecting strength sequence. The segment pairs that meet the connecting conditions are selected from the connecting strength sequence to generate a candidate connecting path set. The entry order of the preceding and following segments in the candidate receiving path set is compared one by one along the time direction to obtain the receiving extension relationship. The receiving extension relationship is then subjected to continuous receiving consistency screening and path integration to generate the control deviation transmission path.

[0013] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the step of analyzing the deviation propagation sequence and expansion trend through a time-series evolution tracing method, determining the starting position and direction of control imbalance, and generating anomaly propagation sequence includes the following specific steps: The temporal evolution tracking method is used to jointly analyze the temporal sequence relationship and the direction of position change of each segment in the control deviation transmission path to obtain the evolutionary correlation relationship. The evolutionary correlation relationship is then combined with the deviation change trend judgment transmission sequence to generate evolution judgment information. Based on the evolution judgment information, the starting position and direction of the control imbalance are determined, and the control deviation propagation path is uniformly arranged in chronological order to generate an anomaly propagation sequence.

[0014] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for performing reverse localization analysis on the anomaly propagation sequence to obtain the location of the anomaly source are as follows: The abnormal propagation sequence is expanded segment by segment to obtain the reverse sequence segment. The start time point identifier, end time point identifier and deviation relationship of each segment are extracted from the reverse sequence segment to generate reverse sequence data. Based on the reverse-order continuation data, the source aggregation of the reverse-order continuation segment pairs is determined, the source aggregation intensity sequence is obtained, and the segment pairs that meet the source characteristics are selected from the source aggregation intensity sequence to generate the source aggregation sequence. The source aggregation sequence is continuously screened, and the segment with the degree of source aggregation is identified as the candidate segment of the anomaly source. The starting point of the candidate segment of the anomaly source is extracted from the candidate segment of the anomaly source to generate the anomaly source location.

[0015] As a preferred embodiment of the fault monitoring method for durian sprinkler irrigation system based on anomaly detection described in this invention, the specific steps for combining the location of the anomaly source with the characteristics of state changes, determining the source and scope of the fault, and generating a fault determination scheme are as follows: The continuity and deviation relationships of points in the anomaly source locations are checked segment by segment in ascending time sequence. Segments that meet the continuity conditions are included in the expansion range to generate impact range data. By combining the location of the anomaly source with the data of its affected area, deviation and change characteristics are obtained. These deviation and change characteristics are then uniformly judged and classified to generate a fault judgment scheme.

[0016] The beneficial effects of this invention are as follows: By uniformly identifying and constructing time-related data for sprinkler irrigation operation, continuous organization and consistent expression of multi-source data are achieved, and matching and consistency analysis of water operation parameters are completed, ensuring stable characterization of operation status; furthermore, by coupling deviation mining and time-series tracking, the deviation relationship of multiple parameters is transformed into a continuously transmitted deviation feature sequence and control deviation transmission path, completing the structured expression of anomalies from point to path; at the same time, by combining time-series evolution tracking and reverse location analysis, the anomaly propagation trend is characterized and the location and impact range of the anomaly source are locked, ultimately forming a fault determination scheme, thereby improving the capabilities of anomaly identification, propagation analysis, and fault location. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0018] Figure 1 A flowchart is generated to establish a consistent sequence for water operation parameters.

[0019] Figure 2 Flowchart for identifying change-related features and generating deviation feature sequences.

[0020] Figure 3 A flowchart is generated to control deviations from the propagation path and abnormal propagation sequences.

[0021] Figure 4 A flowchart is generated for the anomaly source location and fault determination scheme.

[0022] Figure 5 This is a magnified view of the propagation path. Detailed Implementation

[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0024] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0026] Reference Figures 1-5 This is one embodiment of the present invention, which provides a fault monitoring method for a durian sprinkler irrigation system based on anomaly detection, including the following steps: S1. Collect sprinkler irrigation operation data and unify the time and source identifiers of the sprinkler irrigation operation data to generate an operation access sequence.

[0027] S1.1 Collect sprinkler irrigation operation data, read the sprinkler irrigation operation data item by item and sort it by time to generate an initial record set.

[0028] It should be noted that the sprinkler irrigation operation data is extracted item by item, and the collection time, source location, and parameter content corresponding to each sprinkler irrigation operation data are retained; the collection time in the sprinkler irrigation operation data is uniformly converted, and records with missing time, duplicate time, and conflicting content are removed; then all sprinkler irrigation operation data are arranged in order from earliest to latest according to the collection time, and sprinkler irrigation operation data with the same time are further sorted according to the source location; the sorted sprinkler irrigation operation data are continuously summarized in chronological order to generate an initial record set.

[0029] S1.2 Extract source identifier data from the initial record set, combine and mark the source identifier data to obtain the source identifier record set, and perform unified benchmark calibration and time alignment on the collection time of the source identifier record set to generate a unified time record set.

[0030] It should be explained that the process involves reading the collection location, collection channel, and record source information for each record from the initial record set, and then concatenating these information in a preset order to form source identifier data for each record. The source identifier data is then bound to the corresponding record content according to the collection time within the same record, generating a source identifier record set. Subsequently, a unified time base is selected, and the collection times in the source identifier record set are converted one by one. For collection times with time offsets, offset correction is performed; for collection times with inconsistent formats, format unification is performed; and the temporal relationship between records is checked according to the corrected collection times. Finally, the source identifier record set is rearranged according to the unified collection times to generate a time-unified record set.

[0031] It should also be noted that sequential splicing refers to a processing method that combines the collection location, collection channel, and record source information in a unified manner according to a pre-set field arrangement rule. In specific settings, the field priority is determined according to the data structure, such as location first, then channel, and then source, and the field order and format are fixed. During the execution process, the corresponding field content of each record is extracted item by item, and they are connected in the predetermined order to generate a unique identifier, thereby obtaining source identifier data with consistent structure and comparable characteristics.

[0032] S1.3. Arrange the unified time record set continuously according to the time sequence to obtain the access associated record set, and integrate and organize the access associated record set in a unified manner to generate the running access sequence.

[0033] It should be noted that each record in the unified time record set is arranged sequentially from front to back according to the collection time, and records with the same collection time are sorted according to the source identifier data to determine the position of each record in the overall time chain and obtain the access associated record set; the time sequence relationship and source identifier data correspondence of adjacent records in the access associated record set are checked one by one, and records with continuous time and clear source identifier data are merged and organized in chronological order; all records are uniformly integrated and organized in sequence according to the organized sequence to generate the running access sequence.

[0034] It should also be noted that the operation access sequence is used to continuously organize and standardize the sprinkler irrigation operation data according to a unified time sequence and source identification data, so that each record forms a clear sequential relationship in the time dimension and maintains consistent identification in the source dimension. This provides continuous and traceable basic data support for subsequent time correlation construction, water operation parameter extraction and consistency analysis, ensuring the continuity and accuracy of subsequent analysis processes.

[0035] S2. Construct time correlations for the operational access sequence, obtain water operation parameters, and perform time location matching and consistency analysis on the water operation parameters to generate a consistent operational sequence.

[0036] S2.1. Group and expand the running access sequence according to the source identifier data, obtain the location number mapping value, and merge and filter the location number mapping value to generate a location candidate record set.

[0037] It should be noted that each record in the operational access sequence is split item by item according to the source identification data, and the corresponding collection location field, channel field, and associated marker field in the source identification data are extracted one by one to determine the location pointer corresponding to each record; records with the same location pointer are classified and summarized according to the extracted content to obtain the location number mapping value; the location number mapping value is checked item by item to see if the location number is duplicated, if the location pointer is consistent, and if there is a mismatch between the preceding and following records; records with the same location number and clear location pointer are merged and retained, and records with conflicting location numbers or unclear location pointers are filtered out to generate a set of candidate location records.

[0038] It should also be noted that the location number mapping value is used to uniformly encode the collection location and channel information to form a unique location identifier; by merging and mapping records from different sources but corresponding to the same actual location, the data at each time point can be continuously connected and compared under the same spatial reference, thereby providing a stable basis for parameter extraction, time-location matching and deviation identification.

[0039] S2.2 Extract flow rate information, pressure information and soil moisture information from the candidate location record set to construct a candidate set of moisture parameters.

[0040] It should be noted that the location number mapping value and record content of each record are read one by one from the candidate location record set. In each record, flow rate information, pressure information and soil moisture information are searched item by item according to field name and parameter identifier. The found flow rate information, pressure information and soil moisture information are bound to each other, and it is checked whether the flow rate information, pressure information and soil moisture information belong to the same collection time or adjacent receiving time under the same location number mapping value. The records with clear location number mapping values ​​and complete correspondence of flow rate information, pressure information and soil moisture information are summarized and organized to construct a candidate set of moisture parameters.

[0041] S2.3 Match and verify the time sequence and positional correspondence of adjacent records in the candidate set of moisture parameters to obtain the time-position matching sequence, and organize the time-position matching sequence to generate moisture operation parameters.

[0042] It should be noted that each record in the candidate set of moisture parameters is arranged sequentially from front to back according to the collection time, and the location number mapping values ​​of adjacent records are compared one by one to check whether adjacent records meet the requirements of temporal continuity and positional correspondence. For adjacent records that meet the requirements of correct temporal sequence and clear positional correspondence, a succession relationship is established to obtain the time-location matching sequence. According to the succession relationship, the flow rate information, pressure information and soil moisture information in the time-location matching sequence are continuously merged and sorted in order. The parameter records with stable positional correspondence and clear time succession are integrated to generate moisture operation parameters.

[0043] S2.4 Perform consistency calculations on the water operation parameter group to obtain a consistency parameter set. Based on the consistency parameter set, perform continuous arrangement and continuity judgment on the water operation parameters to generate a continuous consistency segment set.

[0044] It should be noted that each parameter record in the water operation parameter group is expanded item by item according to the location number mapping value and the collection time sequence. The flow rate information, pressure information and soil moisture information in each parameter record are checked accordingly. The consistency of adjacent parameter records in terms of the direction of change, the magnitude of change and the temporal continuity is compared to obtain a consistent parameter set. Based on the consistent parameter set, the water operation parameters that meet the consistency judgment conditions are arranged consecutively. The temporal continuity and positional continuity between adjacent water operation parameters are judged one by one. The water operation parameters that are continuously continuous and maintain stable consistency are merged and organized to generate a continuous consistent segment set.

[0045] S2.5. Perform connection and anomaly removal on the continuous consistent segment set to obtain a consistent segment set, and organize the consistent segment set in a unified manner to generate a running consistent sequence.

[0046] It should be noted that, for each continuous consistent segment in the set of continuous consistent segments, each segment is expanded according to the mapping value of the collection time and location number. The time interval, location continuity, and changes in flow rate, pressure, and soil moisture information between adjacent continuous consistent segments are checked. Continuous consistent segments with continuous time, clear location connection, and stable changes are connected and organized to obtain a consistent segment set. Then, each record in the consistent segment set is checked for missing records, abrupt changes, and continuity breaks. Abnormal records that do not meet the continuity conditions are removed. The consistent segment set is then uniformly organized according to the chronological order to generate an operational consistent sequence.

[0047] It should also be noted that the consistent sequence is used to continuously and consistently express water operation parameters in the time and location dimensions, so that flow information, pressure information and soil moisture information form a stable and continuous relationship between adjacent time points, thereby providing a reliable and continuous basis for subsequent change comparison, deviation identification and transmission analysis, and ensuring that the anomaly identification process has clear temporal logic and location correlation basis.

[0048] The acceptance criteria refer to the criteria for determining whether adjacent water operation parameters are continuous in time sequence, have consistent location number mapping values, and maintain a continuous relationship in terms of the direction and magnitude of change of flow rate information, pressure information, and soil moisture information.

[0049] S3. Compare the changes in adjacent time positions of the consistent sequence to obtain the change correlation features. Identify the deviation between flow information, pressure information and soil moisture information from the change correlation features. Then, use the coupled deviation mining method to perform time-series transmission and merging of the deviations to generate a deviation feature sequence. S3.1. Expand the points of the consistent sequence in chronological order to generate points at the same location.

[0050] It should be noted that the consistent sequence of operations is expanded one by one in chronological order, and the location number mapping value corresponding to each record is extracted. Records with the same location number mapping value are grouped consecutively according to the collection time. The consistency of the positions of the records before and after are checked to generate the same location point.

[0051] S3.2 Construct neighboring time records based on the same location points, extract moisture change coupling features from the neighboring time records, and perform difference comparison and correlation integration on the moisture change coupling features to generate change correlation features.

[0052] It should be noted that, based on the same location point, two adjacent records are extracted in chronological order of collection time, and adjacent time records are established sequentially. In each group of adjacent time records, flow rate information, pressure information, and soil moisture information are read separately. The direction, magnitude, and sequence of changes before and after each change are calculated to extract the coupling features of moisture changes. The coupling features of moisture changes in each group are compared according to the time sequence at the same location point to check the synchronous changes and corresponding correlations between flow rate information, pressure information, and soil moisture information. The coupling features of moisture changes with consistent direction, clear connection, and stable correlation are sequentially integrated to generate change correlation features.

[0053] It should also be noted that the moisture change coupling feature refers to the associated feature formed by jointly extracting the direction, magnitude and sequence of changes of flow rate information, pressure information and soil moisture information at adjacent times in the same location; it is used to characterize the coordinated change relationship between multiple parameters, identify whether there are consistent changes or mutual deviations, and thus provide a basis and support for deviation identification and abnormal propagation analysis.

[0054] S3.3. Compare the direction and magnitude of change in the change correlation features item by item, identify change items with inconsistent directions, determine the dominant deviation relationship, and generate deviation information.

[0055] It should be noted that each record in the change correlation feature is expanded sequentially in chronological order. The change direction and magnitude of flow rate, pressure, and soil moisture information in each record are read separately. The change direction of flow rate, pressure, and soil moisture information is compared item by item, and the change magnitude of flow rate, pressure, and soil moisture information is compared item by item to identify change items with opposite change directions or significantly inconsistent change magnitudes. The identified change items are then judged in order of the first occurrence of deviation, the duration of deviation, and the strength of deviation impact to determine the dominant deviation relationship and generate deviation information.

[0056] The expressions for the direction and magnitude of change in the change-related features compared item by item are as follows: ; in, For the first The directional consistency judgment value of the associated features of the change; The change in flow rate information is calculated by subtracting the flow rate information from two water operation parameter records at adjacent time points. The pressure information change is calculated by subtracting the pressure information from two water operation parameter records at adjacent time points. The change in soil moisture information is calculated by subtracting the soil moisture information from two water operation parameter records at adjacent time locations. It is a symbolic function; The first of the change association features arranged in chronological order Index of records.

[0057] S3.4. Using the coupled deviation mining method, check the consistency of the same location points, the consistency of deviation relationships, and the change continuity in the deviation situation item by item along the time direction, and generate a set of propagating deviation segments.

[0058] It should be noted that the coupled deviation mining method is adopted, and the deviation cases are expanded one by one in chronological order. The location number mapping value and dominant deviation relationship corresponding to each deviation case are read. The location number mapping values ​​of adjacent deviation cases are compared one by one to see if they maintain the same location point. The dominant deviation relationship in adjacent deviation cases is checked to see if it is consistent. The start and end of the change in adjacent deviation cases are judged to be continuous. Deviation cases that simultaneously satisfy the consistency of the same location point, the consistency of the deviation relationship, and the continuity of the change are connected to generate a set of progressive deviation segments.

[0059] It should also be noted that the coupled deviation mining method reads the flow rate, pressure, and soil moisture information in the deviation data sequentially at the same location point. First, it compares the direction and magnitude of change of the three data synchronously to determine whether there is a coordinated deviation or a mutually divergent relationship. Then, it continuously compares the deviation data at adjacent time points to check the continuity and transformation of the dominant deviation relationship over time. Subsequently, based on the consistency, duration, and continuity of the deviation relationship, the deviation data that meet the continuous transmission condition are connected sequentially, thereby completing the identification of the multi-parameter coupled deviation process and the construction of the transmission structure.

[0060] It should also be noted that the continuous transmission condition refers to the criteria for determining whether adjacent deviations are consecutive in time sequence, have consistent location number mapping values, and maintain a continuous relationship between the dominant deviation relationship and the direction of change between the preceding and following records.

[0061] S3.5 Perform time continuity and deviation stability checks on the propagation deviation segment set, obtain the propagation check segment set, and segment and organize the broken segments in the propagation check segment set to generate a deviation feature sequence.

[0062] It should be noted that the progressive deviation segment set is expanded segment by segment in chronological order, and the continuity between the start time and end time of each segment is checked. The dominant deviation relationship, direction of change and magnitude of change within each segment are compared to ensure consistency. This completes the verification of time continuity and deviation stability, and yields the progressive verification segment set. The broken segments in the progressive verification segment set are split according to the front and rear boundaries of the break position, and the split continuous segments are rearranged and ordered in chronological order to generate the deviation feature sequence.

[0063] It should also be noted that the deviation feature sequence is used to continuously express and structure the deviation in the time dimension, so that the deviation relationship between flow information, pressure information and soil moisture information forms a traceable transmission chain according to the order of occurrence, thereby providing a clear time evolution basis and deviation correlation basis for subsequent path construction, anomaly propagation analysis and anomaly source location.

[0064] S4. Perform time-series tracking of the deviation feature sequence to obtain the control deviation propagation path, and analyze the deviation propagation order and expansion trend through time-series evolution tracking methods to determine the starting position and direction of control imbalance and generate anomaly propagation sequence.

[0065] S4.1. Expand the deviation feature sequence segment by segment to obtain the preceding and following connecting segments, extract the deviation segment feature information from the preceding and following connecting segments, determine the connection strength of the deviation segment feature information, obtain the connection strength sequence, and select the segment pairs that meet the connection conditions from the connection strength sequence to generate a candidate connection path set.

[0066] It should be explained that the deviation feature sequence is expanded segment by segment in chronological order. Each segment is paired with its immediate successor to obtain the preceding and following segments. Deviation segment feature data is extracted from each of these segments to form deviation segment feature information. The termination time of the preceding segment is compared with the start time of the following segment to determine if the time interval is continuous. The mapping values ​​of the end positions of the preceding and following segments are checked against the mapping values ​​of the beginning positions of the following segments to determine if the positional continuity is clear. The dominant deviation relationship, direction of change, and magnitude of change in the two segments are compared to determine if the deviation continuation is consistent. The strength of the deviation segment feature information is determined based on the degree of temporal continuity, positional continuity, and deviation continuation. Segments with clear continuity are arranged in chronological order to obtain a continuity strength sequence. Finally, segments that are temporally continuous, have clear positional continuity, and maintain a continuous dominant deviation relationship are selected from the continuity strength sequence and merged according to the preceding and following relationships to generate a candidate continuity path set.

[0067] S4.2. Compare the entry order of the preceding and following segments in the candidate receiving path set along the time direction to obtain the receiving extension relationship, and perform continuous receiving consistency screening and path integration on the receiving extension relationship to generate the control deviation transmission path.

[0068] It should be explained that, along the time direction, each group of preceding and following segments in the candidate receiving path set is expanded item by item. The end time of the preceding segment is compared with the start time of the following segment to determine whether the entry order of the preceding and following segments conforms to the chronological relationship. The mapping value of the end position number of the preceding segment is matched with the mapping value of the front position number of the following segment. Combined with the dominant deviation relationship, change direction and change magnitude in the preceding and following segments, it is determined whether the preceding and following segments have the conditions for continuous extension and the receiving extension relationship is obtained. The receiving extension relationship is checked one by one in chronological order. Receiving extension relationships with reversed time, misconnected positions and inconsistent deviation continuation are screened out. The receiving extension relationships that meet the continuous receiving consistency screening conditions are merged and integrated according to the order of preceding and following connections to generate the control deviation transmission path.

[0069] It should also be noted that the continuous extension condition refers to the criteria for determining whether adjacent segments are connected in time sequence, have corresponding position number mapping values, and maintain a consistent dominant deviation relationship and direction of change.

[0070] S4.3. Using the time-series evolution tracking method, the temporal sequence relationship and the direction of position change of each segment in the control deviation transmission path are jointly analyzed to obtain the evolutionary correlation relationship. The evolutionary correlation relationship is then combined with the deviation change trend judgment transmission sequence to generate evolution judgment information.

[0071] It should be noted that the temporal evolution tracing method is adopted to unfold each segment in the control deviation transmission path in chronological order. First, the deviation evolution characteristic data corresponding to each segment are read item by item, and the temporal relationship between two adjacent segments and whether the position change direction maintains continuous extension are compared. The temporal relationship and the position change direction are combined to determine the forward and backward transmission association, directional continuity association and cross-position expansion association between each segment, and to obtain the evolutionary association relationship. The deviation change trend in the evolutionary association relationship is determined item by item, and combined with the forward and backward transmission order of each segment in the control deviation transmission path, the information is uniformly organized to generate evolution determination information.

[0072] It should also be noted that the temporal evolution tracking method reads each segment of the control deviation transmission path in chronological order, extracts the deviation evolution feature data corresponding to each segment, compares the temporal relationship between adjacent segments to determine whether the time continuity is continuous, and verifies the mapping value of the position number between adjacent segments with the direction of change to determine whether the change continues or shifts in the same direction. Subsequently, based on the continuity of the dominant deviation relationship and the magnitude of change, the deviation change process is connected and the relationship is calibrated segment by segment, thereby realizing the identification and continuous tracking of the evolution path of the deviation in time and location.

[0073] Evolutionary determination information is the result of a comprehensive judgment on the transmission sequence, direction of continuation, and expansion trend between segments based on the continuous relationship of deviation in time sequence and location changes. It is used to characterize the evolutionary path and development trend of deviation propagation process, and to provide a basis for determining the starting position and propagation range of control imbalance.

[0074] S4.4 Determine the starting position and direction of control imbalance based on the evolution judgment information, and uniformly arrange the control deviation transmission path in chronological order to generate an anomaly propagation sequence.

[0075] It should be noted that, based on the evolution judgment information, each segment is examined sequentially according to the order of transmission, changes in position number mapping values, and the continuation of dominant deviation relationships. The position number mapping value of the first continuous deviation transmission is determined as the starting position of the control imbalance. The starting direction of the control imbalance is determined based on the direction of point change between the segment where the starting position of the control imbalance is located and the next segment. Each segment in the control deviation transmission path is arranged sequentially according to time, and the starting position of the control imbalance, the starting direction of the control imbalance, and the connection relationship between each segment are organized together to generate an abnormal propagation sequence.

[0076] like Figure 5As shown, under repeated simulation conditions, there are differences in the characterization results of the anomaly propagation process. The control deviation transmission path identified by the experimental group is closer to the actual propagation path as a whole, with a smaller path deviation. This indicates that after forming the deviation feature sequence, the present invention can obtain the control deviation transmission path through time-series tracing and analyze the deviation transmission order and expansion trend by combining the time-series evolution tracing method, making the generated anomaly propagation sequence more consistent with the actual propagation process. The enlarged local image further shows that in the interval where the difference is more obvious, the experimental group has a better ability to identify the key propagation segment than the control group, which better reflects the unified characterization of the anomaly propagation path, propagation order and expansion trend of the present invention, and provides a clearer path basis for the determination of the anomaly source location and the analysis of the impact range.

[0077] The control group used a sequential comparison method based on a single time series, which independently detected changes in flow, pressure and soil moisture information. It judged anomalies based solely on the temporal relationship, without performing location mapping correlation, coupling deviation analysis and propagation path continuation modeling. Therefore, it could only obtain discrete anomaly points, which was difficult to characterize the continuous propagation process, resulting in large path deviation and weak source localization ability.

[0078] It should also be noted that the anomaly propagation sequence is used to continuously express the expansion process of control imbalance in the time and location dimensions, so that the transmission path of the deviation relationship along the mapping values ​​of each location number is clearly presented in chronological order, thereby providing a complete propagation trajectory basis and evolution process support for subsequent anomaly source location and impact range determination.

[0079] S5. Perform reverse localization analysis on the abnormal propagation sequence to obtain the location of the abnormal source, combine the location of the abnormal source with the state change characteristics, determine the source of the fault and the scope of its impact, and generate a fault determination scheme.

[0080] S5.1. Expand the abnormal propagation sequence segment by segment to obtain the reverse sequence segment. Extract the start time point identifier, end time point identifier and deviation relationship of each segment from the reverse sequence segment to generate reverse sequence data.

[0081] It should be noted that the abnormal propagation sequence is expanded segment by segment in reverse chronological order. The next segment of the abnormal propagation sequence is paired with the preceding segment that is immediately adjacent in time to obtain the reverse sequence continuation segment. From each group of reverse sequence continuation segments, the reverse sequence continuation feature data corresponding to the preceding and following segments are extracted item by item. The connection between the start time and the end time, the continuation of the location identifiers, and the continuation of the deviation relationship are checked item by item. The checked start time, end time, location identifiers, and deviation relationships are sorted in reverse sequence continuation order to generate reverse sequence continuation data.

[0082] It should also be noted that the reverse-order data is structured information formed by extracting and organizing the time connection, location correspondence and deviation relationship between adjacent segments after unfolding the abnormal propagation sequence in reverse chronological order. It is used to trace the source of deviation from the propagation result in reverse, strengthen the source aggregation and judgment process, and thus provide a reliable basis for identifying the location of the abnormal source and confirming the propagation starting point.

[0083] S5.2. Based on the reverse-order continuation data, perform source aggregation determination on the pairs of reverse-order continuation segments, obtain the source aggregation intensity sequence, and select the segment pairs that meet the source characteristics from the source aggregation intensity sequence to generate the source aggregation sequence.

[0084] It should be noted that, based on the reverse-order continuation data, the pairs of consecutive reverse-order continuation segments are expanded group by group according to the reverse-order continuation sequence, and the source aggregation judgment features corresponding to each group of consecutive reverse-order continuation segment pairs are extracted. The time continuity, point concentration, and deviation relationship convergence in the source aggregation judgment features are compared and comprehensively judged item by item to determine the source aggregation degree of each segment pair and obtain the source aggregation intensity sequence. Each group of segment pairs in the source aggregation intensity sequence is checked item by item in reverse order. First, the interval between the end time of the previous segment and the start time of the next segment in each group of segment pairs is compared to see if it is short and maintains continuity. Next, check whether the corresponding point markers of the two segments converge towards the same point or towards adjacent points. Then, compare whether the dominant deviation relationship, direction of change, and magnitude of change in the two segments gradually change from scattered changes to concentrated consistency. Segment pairs with continuous time connection, clear point concentration, and convergent deviation relationship are retained as segment pairs that meet the source characteristics. Segment pairs with obvious time jumps, scattered points without a concentrated trend, or divergent deviation relationship are eliminated. Finally, the retained segment pairs are connected and sorted in reverse order to generate the source aggregation sequence.

[0085] It should also be noted that the source aggregation intensity sequence is an ordered sequence formed by comprehensively evaluating the degree of temporal connection, the degree of concentration of points, and the degree of convergence of deviation relationships of the reverse succession segments; it is used to measure the strength of each segment converging towards the same source, thereby screening out the most likely candidate regions of anomaly sources and providing a basis for determining the location of anomaly sources.

[0086] S5.3. Perform continuous screening on the source aggregation sequence, and determine the segment with the degree of source aggregation as the candidate segment of the abnormal source. Extract the starting point of the candidate segment of the abnormal source from the candidate segment of the abnormal source to generate the abnormal source location.

[0087] It should be noted that the source aggregation sequence is expanded segment by segment in reverse order, and each segment is checked to see if the temporal connection, point connection, and deviation continuity of adjacent segments are maintained. Segments with continuous correspondence are then selected for continuous connection. After continuous connection selection, each segment is compared according to the degree of source aggregation, and the segment with the highest degree of source aggregation is identified as the candidate segment for anomaly source. The corresponding starting point is extracted from the candidate segment for anomaly source, and the starting points of multiple candidate segments for anomaly source are merged and verified to generate the anomaly source location.

[0088] S5.4. Check the continuity of the connection and deviation relationships of the points in the anomaly source location segment by segment in ascending time sequence, and include the segments that meet the connection conditions into the expansion range to generate the impact range data.

[0089] It should be noted that, following the chronological order, each segment of the record corresponding to the location of the anomaly source is expanded, and the connection between the current segment and the next segment is checked item by item to determine whether the location marker of the next segment corresponds to the extension direction of the current segment. The continuity of the deviation relationship between the current segment and the next segment is checked item by item to compare whether the dominant deviation relationship, change direction and change magnitude of the two segments remain continuous. Segments that simultaneously satisfy the continuity of the location connection relationship and the stability of the deviation relationship are included in the expansion range, and the segments included in the expansion range are merged and organized in chronological order to generate the impact range data.

[0090] S5.5 Combine the data on the location of the anomaly source with the range of influence to obtain the deviation change characteristics, and perform unified judgment and hierarchical organization of the deviation change characteristics to generate a fault judgment scheme.

[0091] It should be explained that the data on the location of the anomaly source and the scope of influence are matched item by item according to the time sequence and the correspondence between the points. The starting point in the location of the anomaly source is extracted and the extended point in the scope of influence is extracted. The deviation change characteristics are obtained by combining the dominant deviation relationship, change direction and change magnitude corresponding to each point. The deviation change characteristics are checked item by item for the deviation starting position, the size of the extended scope, the intensity of the deviation, and the degree of change concentration. A unified judgment is made based on the check content. The deviation change characteristics are classified and organized according to the degree of deviation and the degree of influence. The judgment content after classification is summarized in order to generate a fault judgment plan.

[0092] In summary, this invention achieves continuous organization and consistent expression of multi-source data by uniformly identifying and temporally associating sprinkler irrigation operation data, and completing the matching and consistency analysis of water operation parameters to ensure stable operation status characterization. Furthermore, by coupling deviation mining and temporal follow-up tracking, the multi-parameter deviation relationship is transformed into a continuously transmitted deviation feature sequence and control deviation transmission path, completing the structured expression of anomalies from point to path. At the same time, by combining temporal evolution tracking and reverse location analysis, the anomaly propagation trend is characterized and the location and impact range of the anomaly source are locked, ultimately forming a fault determination scheme, thereby improving the capabilities of anomaly identification, propagation analysis, and fault location.

[0093] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fault monitoring method for a durian sprinkler irrigation system based on anomaly detection, characterized in that, include: Collect sprinkler irrigation operation data and unify the time and source identifiers of the sprinkler irrigation operation data to generate an operation access sequence; Time correlation is constructed for the operational access sequence to obtain water operation parameters, and time location matching and consistency analysis are performed on the water operation parameters to generate a consistent operational sequence; The changes in adjacent time positions of the consistent sequence are compared to obtain the change correlation features. From the change correlation features, the deviation between flow rate information, pressure information and soil moisture information is identified. The coupled deviation mining method is used to merge the deviations in time sequence to generate a deviation feature sequence. The deviation feature sequence is tracked over time to obtain the control deviation propagation path. The propagation order and expansion trend of the deviation are analyzed by the time-series evolution tracking method to determine the starting position and direction of control imbalance and generate anomaly propagation sequence. The abnormal propagation sequence is reversed and analyzed to obtain the location of the abnormal source. The location of the abnormal source is combined with the characteristics of state change to determine the source and scope of the fault and generate a fault determination scheme.

2. The method for fault monitoring of durian sprinkler irrigation system based on anomaly detection as described in claim 1, characterized in that, The specific steps for collecting sprinkler irrigation operation data, unifying the time and source identifiers of the sprinkler irrigation operation data, and generating an operation access sequence are as follows: Collect sprinkler irrigation operation data, read the sprinkler irrigation operation data item by item and sort it by time to generate an initial record set; Source identifier data is extracted from the initial record set, the source identifier data is combined and marked to obtain the source identifier record set, and the collection time of the source identifier record set is uniformly calibrated and time-aligned to generate a time-unified record set; The unified time record set is arranged continuously in chronological order to obtain the access-related record set. The access-related record set is then integrated and organized in a unified manner to generate the running access sequence.

3. The method for fault monitoring of durian sprinkler irrigation system based on anomaly detection as described in claim 2, characterized in that, The specific steps for constructing time-related correlations of the operational access sequences to obtain water operation parameters are as follows: The operational access sequences are grouped and expanded according to the source identifier data to obtain the location number mapping value. The location number mapping value is then merged and filtered to generate a set of candidate location records. Flow rate, pressure and soil moisture information are extracted from the candidate location record set to construct a candidate set of moisture parameters; The temporal order and positional correspondence of adjacent records in the candidate set of moisture parameters are matched and verified to obtain the time-position matching sequence. The time-position matching sequence is then combined and organized to generate moisture operation parameters.

4. The method for fault monitoring of durian sprinkler irrigation system based on anomaly detection as described in claim 1, characterized in that, The specific steps for performing time-location matching and consistency analysis on water operation parameters to generate a consistent operation sequence are as follows: Consistency calculations are performed on the water operation parameter group to obtain a consistency parameter set. Based on the consistency parameter set, the water operation parameters are continuously arranged and their continuity is determined to generate a continuous consistency segment set. The continuous consistent segment set is connected, sorted, and anomaly removed to obtain a consistent segment sorting set. The consistent segment sorting set is then uniformly organized to generate a running consistent sequence.

5. The method for fault monitoring of durian sprinkler irrigation system based on anomaly detection as described in claim 4, characterized in that, The specific steps for comparing adjacent time position changes in the consistent sequence to obtain change correlation features are as follows: Expand the points of the consistent running sequence according to time order to generate points at the same location; Construct neighboring time records based on the same location points, extract moisture change coupling features from the neighboring time records, and perform difference comparison and correlation integration on the moisture change coupling features to generate change correlation features.

6. The method for fault monitoring of a durian sprinkler irrigation system based on anomaly detection as described in claim 1, characterized in that, The deviations between flow rate, pressure, and soil moisture information are identified from the change correlation features. A coupled deviation mining method is then used to perform time-series propagation and merging of the deviations to generate a deviation feature sequence. The specific steps are as follows: By comparing the direction and magnitude of change in the change correlation features item by item, we can identify change items with inconsistent directions, determine the dominant deviation relationship, and generate deviation information. The coupled deviation mining method is used to check the consistency of the same location, the consistency of the deviation relationship, and the change continuity of the deviation along the time direction, and to generate a set of progressive deviation segments. The time continuity and deviation stability of the propagation deviation segment set are checked to obtain the propagation check segment set. The broken segments in the propagation check segment set are segmented and organized to generate the deviation feature sequence.

7. The method for fault monitoring of a durian sprinkler irrigation system based on anomaly detection as described in claim 6, characterized in that, The specific steps for performing time-series tracing of the deviation feature sequence to obtain the control deviation propagation path are as follows: The deviation feature sequence is expanded segment by segment to obtain the preceding and following connecting segments. The deviation segment feature information is extracted from the preceding and following connecting segments. The connecting strength of the deviation segment feature information is determined to obtain the connecting strength sequence. The segment pairs that meet the connecting conditions are selected from the connecting strength sequence to generate a candidate connecting path set. The entry order of the preceding and following segments in the candidate receiving path set is compared one by one along the time direction to obtain the receiving extension relationship. The receiving extension relationship is then subjected to continuous receiving consistency screening and path integration to generate the control deviation transmission path.

8. The method for fault monitoring of a durian sprinkler irrigation system based on anomaly detection as described in claim 1, characterized in that, The method of using time-series evolution tracing analyzes the deviation from the propagation order and expansion trend, determines the starting position and direction of the control imbalance, and generates an abnormal propagation sequence. The specific steps are as follows: The temporal evolution tracking method is used to jointly analyze the temporal sequence relationship and the direction of position change of each segment in the control deviation transmission path to obtain the evolutionary correlation relationship. The evolutionary correlation relationship is then combined with the deviation change trend judgment transmission sequence to generate evolution judgment information. Based on the evolution judgment information, the starting position and direction of the control imbalance are determined, and the control deviation propagation path is uniformly arranged in chronological order to generate an anomaly propagation sequence.

9. The method for fault monitoring of a durian sprinkler irrigation system based on anomaly detection as described in claim 8, characterized in that, The specific steps for performing reverse localization analysis on the abnormal propagation sequence to obtain the location of the abnormal source are as follows: The abnormal propagation sequence is expanded segment by segment to obtain the reverse sequence segment. The start time point identifier, end time point identifier and deviation relationship of each segment are extracted from the reverse sequence segment to generate reverse sequence data. Based on the reverse-order continuation data, the source aggregation of the reverse-order continuation segment pairs is determined, the source aggregation intensity sequence is obtained, and the segment pairs that meet the source characteristics are selected from the source aggregation intensity sequence to generate the source aggregation sequence. The source aggregation sequence is continuously screened, and the segment with the degree of source aggregation is identified as the candidate segment of the anomaly source. The starting point of the candidate segment of the anomaly source is extracted from the candidate segment of the anomaly source to generate the anomaly source location.

10. The method for fault monitoring of a durian sprinkler irrigation system based on anomaly detection as described in claim 9, characterized in that, The steps for combining the location of the anomaly source with the characteristics of state changes, determining the source and scope of the fault, and generating a fault determination scheme are as follows: The continuity and deviation relationships of points in the anomaly source locations are checked segment by segment in ascending time sequence. Segments that meet the continuity conditions are included in the expansion range to generate impact range data. By combining the location of the anomaly source with the data of its affected area, deviation and change characteristics are obtained. These deviation and change characteristics are then uniformly judged and classified to generate a fault judgment scheme.

Images