Underwater Intelligent Pipeline Fault Detection Method and System Based on Sensor Breakpoint Detection

Abstract

The present invention discloses an underwater intelligent pipe network fault detection method and system based on sensor breakpoint detection, belonging to the technical field of fault detection. The key points of its technical solution include: determining a faulty sensor and its corresponding extended sensor, and the radiation detection range of the extended sensor includes the central detection range of the faulty sensor; determining the untransmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extended sensor at the current moment; determining the fault type of the underwater intelligent pipe network according to the untransmitted data and the fault association rule, and the fault association rule is determined according to the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipe network. The present invention accurately estimates the data of the faulty sensor based on the data of the extended sensor, determines the fault association rule based on the historical fault conditions, and obtains the accurate fault type of the pipe network through the fault association model, improving the accuracy of fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and more particularly to an underwater intelligent pipe network fault detection method and system based on sensor breakpoint detection. Background Art

[0002] Underwater pipe networks are widely used in many fields such as urban water supply, drainage, industrial water conveyance, oil conveyance, and gas conveyance. Their safe and stable operation is crucial. However, due to the complex underwater environment, underwater pipe networks face various fault risks, which makes the development of underwater pipe network fault detection technology an inevitable requirement.

[0003] For example, Chinese Patent Application No. CN114353886A discloses a method and system for detecting fault points in an urban drainage pipe network. By calculating the difference between the flow mean curve during the period to be detected and the historical flow mean curve of the corresponding period, the fault area and the location of the fault point are determined based on the obtained flow error curve.

[0004] However, this method depends on the representativeness of historical flow data. If the operating conditions, water usage patterns, or characteristics of the conveying medium of the underwater pipe network suddenly change, the historical data cannot accurately reflect the current situation. Therefore, there are deficiencies in the prior art. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the purpose of the present invention is to provide an underwater intelligent pipe network fault detection method and system based on sensor breakpoint detection, which determines the type of pipe network fault through the transmitted data of the extended sensors and the fault correlation rules, so as to solve the problem that historical data cannot accurately reflect the current situation during sudden faults, and thus the type of fault cannot be determined.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] The present invention provides an underwater intelligent pipe network fault detection method based on sensor breakpoint detection. A plurality of sensors are arranged in the underwater intelligent pipe network, and each sensor corresponds to a central detection range and a radiation detection range. The underwater intelligent pipe network fault detection method includes:

[0008] Determine the faulty sensor and its corresponding extended sensor. The faulty sensor is a sensor whose data transmission is interrupted at the current moment, and the radiation detection range of the extended sensor includes the central detection range of the faulty sensor;

[0009] Determine the data not transmitted by the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extended sensor at the current moment;

[0010] Determine the fault type of the underwater intelligent pipeline network according to the untransmitted data and the fault association rule, where the fault association rule is determined according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater intelligent pipeline network.

[0011] As a further improvement of the present invention, determining the untransmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the outlying sensor at the current moment includes:

[0012] Obtain the first predicted data and the second predicted data respectively according to the historical transmission data of the faulty sensor and the transmission data of the outlying sensor at the current moment;

[0013] Determine the fault type of the faulty sensor according to the historical transmission data of the faulty sensor and the transmission data of the outlying sensor at the current moment, where the fault type of the faulty sensor includes sudden faults and progressive faults;

[0014] Obtain the untransmitted data according to the first predicted data, the second predicted data and the fault type.

[0015] As a further improvement of the present invention, the obtaining the first predicted data and the second predicted data respectively according to the historical transmission data of the faulty sensor and the transmission data of the outlying sensor at the current moment includes:

[0016] Obtain the first predicted data according to the historical transmission data of the faulty sensor and the time series model;

[0017] Obtain the second predicted data according to the transmission data of the outlying sensor at the current moment and the interpolation algorithm.

[0018] As a further improvement of the present invention, obtaining the second predicted data according to the transmission data of the outlying sensor at the current moment and the interpolation algorithm includes:

[0019] Determine the triangular network structure according to the position of the outlying sensor;

[0020] Determine the triangle to which the faulty sensor belongs according to the position of the faulty sensor and the triangular network structure;

[0021] Determine multiple sub-triangles according to the position of the faulty sensor and the triangle;

[0022] Obtain the second predicted data according to the areas of the multiple sub-triangles and the transmission data of the outlying sensor at the current moment.

[0023] As a further improvement of the present invention, the fault association rule is determined according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater intelligent pipe network, and includes:

[0024] Obtain a transaction data set according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater intelligent pipe network;

[0025] Determine preliminary association rules based on the item sets;

[0026] Obtain the fault association rule according to the preset evaluation index and the preliminary association rule.

[0027] As a further improvement of the present invention, obtaining a transaction data set according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater intelligent pipe network includes:

[0028] Determine the factors affecting the operation state of the underwater intelligent pipe network according to the historical transmission data of the multiple sensors;

[0029] Determine dynamic thresholds according to the factors and the historical transmission data of the multiple sensors;

[0030] Perform dynamic discretization processing on the historical transmission data of the multiple sensors according to the dynamic thresholds to obtain discretized data;

[0031] Organize the discretized data into the transaction data set according to the historical fault conditions of the underwater intelligent pipe network.

[0032] As a further improvement of the present invention, determining preliminary association rules based on the transaction data set includes:

[0033] Gradually generate multiple candidate item sets based on each item set in the transaction data set;

[0034] Calculate the support degrees of the multiple candidate item sets, and determine multiple frequent item sets according to the support degrees;

[0035] Determine preliminary association rules according to the frequent item sets.

[0036] As a further improvement of the present invention, gradually generating multiple candidate item sets based on each item set in the transaction data set includes:

[0037] Perform transaction compression processing on each item set in the transaction data set to obtain multiple compressed transactions;

[0038] Gradually generate the multiple candidate item sets based on the multiple compressed transactions.

[0039] As a further improvement of the present invention, the preset evaluation indicators include confidence, lift, and conviction.

[0040] The present invention provides an underwater intelligent pipeline network fault detection system based on sensor breakpoint detection. A plurality of sensors are arranged in the underwater intelligent pipeline network, and each sensor corresponds to a central detection range and a radiation detection range. The underwater intelligent pipeline network fault detection system includes:

[0041] A positioning module: determining a faulty sensor and its corresponding extension sensors, where the faulty sensor is a sensor whose data transmission is interrupted at the current moment, and the radiation detection range of the extension sensor includes the central detection range of the faulty sensor;

[0042] A calculation module: determining the untransmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment;

[0043] A detection module: determining a fault association rule according to the historical transmission data of the plurality of sensors and the historical fault conditions of the underwater intelligent pipeline network, and determining the fault type of the underwater intelligent pipeline network according to the untransmitted data and the fault association rule.

[0044] When a breakpoint occurs in the sensor, the present invention effectively restores the untransmitted data of the faulty sensor at the current moment based on the historical transmission data of the faulty sensor and the current transmission data of the extension sensor, improves the data quality, and accurately locates the fault type in combination with the fault association rule, thereby improving the detection efficiency. Description of the Drawings

[0045] Figure 1 It is a flowchart of the method steps of the present invention.

[0046] Figure 2 It is a schematic diagram of the central detection range and the radiation detection range.

[0047] Figure 3 It is a schematic diagram of a triangular mesh structure.

[0048] Figure 4 It is a schematic diagram of the system structure of the present invention. Detailed Embodiments

[0049] The technical solution of the present invention will be described in detail below through the drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention.

[0050] The term "and / or" in the following text is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. Additionally, the character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0051] As Figure 1 shown, an embodiment of the present application provides an underwater intelligent pipeline network fault detection method based on sensor breakpoint detection, including:

[0052] Determine the faulty sensor and its corresponding extended sensors, where the faulty sensor is a sensor whose data transmission is interrupted at the current moment;

[0053] Determine the data not transmitted by the faulty sensor at the current moment based on the historical transmission data of the faulty sensor and the transmission data of the extended sensors at the current moment;

[0054] Determine the fault type of the underwater intelligent pipeline network according to the untransmitted data and the fault association rule, where the fault association rule is determined based on the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipeline network.

[0055] Among them, a sensor whose data transmission is interrupted at the current moment can also be called a sensor with a breakpoint; each sensor transmits the detected data to the server for subsequent data analysis; the types of sensors include: pressure sensors, flow sensors, water quality sensors, etc.; the fault types of the underwater intelligent pipeline network include leakage faults, blockage faults, corrosion faults, etc.

[0056] And as Figure 2 shown, multiple sensors are provided in the underwater intelligent pipeline network, and each sensor corresponds to a central detection range and a radiation detection range. The central detection range refers to the core area where the sensor can obtain specific parameter information of the pipeline network with high accuracy. For example, the central detection range of a pressure sensor is usually within 0.5 meters around the sensor installation point. The radiation detection range refers to a relatively large area extending from the sensor to the surrounding with the central detection range included, and this area enables the sensor to conduct a more extensive monitoring and preliminary judgment on the pipeline network parameters. For example, the radiation detection range of a pressure sensor can reach an area with a radius of 1 meter centered on the installation point. Figure 2 The solid circle represents the central detection range, and the dashed circle represents the radiation detection range. The central detection range of each sensor can be included in the radiation detection ranges of multiple sensors of the same type. Therefore, these multiple sensors are called the extended sensors corresponding to this sensor. For example, the extended sensors of sensor A are sensors B, C, D, E, and F. When there is no faulty sensor, the data transmitted by each sensor at each moment is used as the detection data within the corresponding central detection range.

[0057] In this embodiment, when a breakpoint occurs in the sensor, the historical transmission data of the faulty sensor and the current transmission data of the extension sensor are used to effectively restore the untransmitted data of the faulty sensor at the current moment, improve the data quality, and accurately locate the fault type in combination with the fault correlation rules, thereby improving the detection efficiency.

[0058] Furthermore, this embodiment provides a step of determining the untransmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the current transmission data of the extension sensor, including:

[0059] Respectively obtain the first predicted data and the second predicted data according to the historical transmission data of the faulty sensor and the current transmission data of the extension sensor;

[0060] Determine the fault type of the faulty sensor according to the historical transmission data of the faulty sensor and the current transmission data of the extension sensor. The fault types of the faulty sensor include sudden faults and progressive faults;

[0061] Obtain the untransmitted data according to the first predicted data, the second predicted data, and the fault type.

[0062] Exemplarily, if the performance of the pipeline area monitored by the faulty sensor gradually changes due to long-term corrosion, the historical transmission data of the faulty sensor will show a trend of gradual change over time. At this time, the fault type is a progressive fault; if the historical transmission data of the faulty sensor does not show a trend of change over time, and the historical transmission data is within the normal range, while the current transmission data of the extension sensor has a sudden change compared with the transmission data of the previous moment, the fault type is a sudden fault at this time. For example, when the underwater pipe network suddenly bursts due to external construction, resulting in a breakpoint in the pressure sensor; if the historical transmission data of the faulty sensor and the current transmission data of the extension sensor do not fall into the above situations, the reason for the breakpoint needs to be considered from the perspective of the sensor itself, such as hardware damage and power failure. This embodiment only considers the sensor breakpoints caused by pipe network faults.

[0063] Furthermore, this embodiment provides a step of respectively obtaining the first predicted data and the second predicted data according to the historical transmission data of the faulty sensor and the current transmission data of the extension sensor, including:

[0064] Obtain the first predicted data according to the historical transmission data of the faulty sensor and the time series model;

[0065] Obtain the second predicted data according to the current transmission data of the extension sensor and the interpolation algorithm.

[0066] Specifically, after obtaining the first prediction data and the second prediction data, different weights can be assigned to the first prediction data and the second prediction data according to the type of fault. That is, when a sudden fault occurs, a larger weight is assigned to the second prediction data, and when a progressive fault occurs, a larger weight is assigned to the first prediction data. Finally, the untransmitted data is obtained through weighted fusion. Preferably, since the historical transmission data of the faulty sensor has a much smaller impact on the current untransmitted data than the current transmission data of the external sensor at the current moment, the weight corresponding to the second prediction data should be greater than 0.8 at this time.

[0067] Based on the first prediction data and the second prediction data, this embodiment can restore relatively accurate untransmitted data under different types of faults, which serves as the basis for subsequent analysis.

[0068] Furthermore, in addition to the above-mentioned weighted fusion method, a step of determining the untransmitted data of the faulty sensor at the current moment based on a hierarchical fusion method can also be adopted. The principle of hierarchical fusion is that first, a method is used for preliminary estimation, and then the result is used as the input of another method for further optimization. Exemplarily, when a progressive fault occurs, the second prediction data and the historical transmission data of the faulty sensor can be jointly input into a time series model, and the result output by the time series model is used as the untransmitted data. This method takes the second prediction data as a feature and inputs it into the time series model to enhance the adaptability of the time series model to complex situations, so as to obtain a more accurate output result.

[0069] Furthermore, this embodiment provides a step of obtaining the second prediction data according to the current transmission data of the external sensor and an interpolation algorithm, including:

[0070] Determine a triangular network structure according to the position of the external sensor;

[0071] Determine the triangle to which the faulty sensor belongs according to the position of the faulty sensor and the triangular network structure;

[0072] Determine multiple sub-triangles according to the position of the faulty sensor and the triangle;

[0073] Obtain the second prediction data according to the areas of the multiple sub-triangles and the current transmission data of the external sensor.

[0074] Exemplarily, such as Figure 3As shown, the extended sensors of sensor A are sensor B, sensor C, sensor D, sensor D, and sensor F. First, for each extended sensor, connect it to the adjacent sensors corresponding to this extended sensor to form a triangular network structure. Among them, the corresponding adjacent sensors are defined as other extended sensors whose radiation detection ranges overlap with that of this extended sensor. Then, determine the triangles containing the faulty sensor. When there are multiple triangles containing the faulty sensor, such as Figure 3 the triangles containing the faulty sensor A in and , calculate the overlapping area of the regions formed by the three extended sensors in the two triangles with the central detection range of sensor A respectively. Take the triangle corresponding to the larger area as the triangle to which the faulty sensor belongs. If the two calculated areas are the same, such as Figure 3 where the regions formed by the three extended sensors in the two triangles can completely cover the central detection range of sensor A, randomly select one of the two triangles as the triangle to which the faulty sensor belongs. In this embodiment, select . After that, connect sensor A to the three vertices of to obtain three sub-triangles , and , and calculate the areas of , , and respectively. According to the ratio of the area of each sub-triangle to the area of , determine the second prediction data . The specific formula is:

[0075] ;

[0076] where , , and represent the areas of , , and respectively, and , , represent the transmission data of the extended sensors B, D, and E at the current moment respectively.

[0077] In this embodiment, by considering the area of a triangle to take into account the overall characteristics within the triangular region, it can better reflect the spatial distribution of data and is more suitable for sensors with irregular distributions. For example, in a local area of an underwater pipe network, due to the complexity of the underwater pipe network, the sensor distribution is uneven. Area ratio interpolation can reasonably allocate the influence of the sensor data at each vertex on the faulty sensor data according to the triangulation. Moreover, when outliers occur at the outlying sensors at the vertices of the triangle due to external environments or hardware aging, etc., the calculation method provided in this embodiment based on area ratio can mitigate the influence brought by the outliers and increase the accuracy of the calculation results compared with the traditional calculation method based on distance to determine weights.

[0078] Exemplarily, taking as an example, when sensor E has an outlier due to electromagnetic interference, if the method based on distance to determine weights is adopted, since sensor E is the closest to the faulty sensor A, the outlier will be assigned a relatively large weight, and the accuracy of the second prediction data will be greatly affected. However, if the method provided in this embodiment is adopted, since has the smallest area, the outlier will be assigned a relatively small weight, thereby mitigating the influence of the outlier.

[0079] Furthermore, this embodiment provides steps for determining a fault association rule based on the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipe network, including:

[0080] Obtaining a transaction dataset based on the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipe network;

[0081] Determining a preliminary association rule based on the item set;

[0082] Obtaining a fault association rule according to the preset evaluation index and the preliminary association rule.

[0083] Furthermore, this embodiment provides steps for obtaining a transaction dataset based on the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipe network, including:

[0084] Determining the factors affecting the operation state of the underwater intelligent pipe network according to the historical transmission data of multiple sensors;

[0085] Determining a dynamic threshold according to the factors and the historical transmission data of multiple sensors;

[0086] Performing dynamic discretization processing on the historical transmission data of multiple sensors according to the dynamic threshold to obtain discretized data;

[0087] Sorting out the discretized data into a transaction dataset according to the historical fault conditions of the underwater intelligent pipe network.

[0088] Among them, the factors affecting the operating state of the underwater pipe network can be seasons or working conditions. For example, in spring and summer, since the domestic water consumption of residents will increase significantly, this makes the underwater pipe network face greater water supply pressure and the flow demand will also increase. Therefore, in spring and summer, the thresholds of the pressure sensor and the flow sensor should be higher than those in autumn and winter; similarly, under high-load working conditions, the thresholds of the pressure sensor and the flow sensor should be higher than those under low-load working conditions.

[0089] Exemplarily, this embodiment provides a method for determining dynamic thresholds when seasons are used as influencing factors. Assume that among multiple sensors, there are multiple pressure sensors and multiple flow sensors. First, for multiple pressure sensors, classify their historical transmission data according to seasons to obtain four subsets of pressure data. Then calculate the mean and standard deviation of each subset. Assume that the mean of one subset is and the standard deviation is , take and as the thresholds corresponding to this season, where and are coefficients determined according to experience. For example, . Repeat this step for the other three subsets to obtain different thresholds corresponding to each season. For multiple flow sensors, repeat the above steps to obtain different thresholds corresponding to each season for the flow sensors.

[0090] The method provided in this embodiment can dynamically adjust appropriate thresholds according to different seasons or working conditions to discretize data, so that the discretized sensor data can more accurately reflect the actual operating state of the current underwater pipe network, providing a more effective data basis for subsequent association rule mining.

[0091] Furthermore, this embodiment provides steps for determining preliminary association rules based on a transaction dataset, including:

[0092] Based on each item set in the transaction dataset, gradually generate multiple candidate item sets;

[0093] Calculate the support degrees of multiple candidate item sets and determine multiple frequent item sets according to the support degrees;

[0094] Determine preliminary association rules according to the frequent item sets.

[0095] Furthermore, this embodiment provides steps for gradually generating multiple candidate item sets based on each item set in the transaction dataset, including:

[0096] Perform transaction compression processing on each item set in the transaction dataset to obtain multiple compressed transactions;

[0097] Generate multiple candidate sets step by step based on multiple compression transactions.

[0098] Exemplarily, when it is necessary to detect faults in an underwater pipe network within a preset range, it is first necessary to establish fault association rules within the preset range. Assume that there are 3 pressure sensors and 3 flow sensors within the preset range, and in this example, the influencing factor is the season. First, obtain the historical transmission data of the 6 sensors and the influencing factor and perform discretization processing based on the influencing factor. Specifically, for the pressure sensors, obtain the threshold corresponding to the current season according to the season, and divide the data of the 3 pressure sensors into multiple intervals based on the threshold. For example, divide the data less than into the "low" interval, divide the data greater than or equal to and less than into the "normal" interval, and divide the data greater than or equal to into the "high" interval. Similarly, for the flow sensors, the data of the 3 flow sensors can be divided into the "low flow", "normal flow", and "high flow" intervals.

[0099] Then, organize the discretized data to obtain a transaction dataset. The transaction dataset includes multiple transactions, and each transaction includes the status of all sensors within a moment and the information on whether a fault has occurred. For example, one of the transactions is [pressure sensor 1 = high, pressure sensor 2 = normal, pressure sensor 3 = low, flow sensor 1 = normal flow, flow sensor 2 = normal flow, flow sensor 3 = low flow, pipe network fault = leakage fault]. Each transaction includes multiple single-item sets. For example, [pressure sensor 1 = high], [pressure sensor 2 = low], [flow sensor 1 = normal flow], etc. are all single-item sets.

[0100] After that, perform transaction compression processing based on each single-item set. Specifically, first calculate the support degree of each single-item set. The support degree refers to the ratio of the number of transactions of the single-item set in the transaction dataset to the total number of transactions. The higher the support degree, the higher the frequency of the single-item set. Then, compare the support degree of each single-item set with the preset support degree, and use the single-item sets with support degrees greater than the preset support degree as preferred single-item sets. Then, perform transaction compression processing on the transaction dataset according to the preferred single-item sets, that is, only retain the transactions including at least one preferred single-item set in the transaction dataset, and call these transactions compressed transactions.

[0101] Next, connect each pair of preferred single-item sets to obtain multiple candidate 2-item sets. For example, if the preferred single-item sets are [Pressure Sensor 1 = High], [Pressure Sensor 2 = Low], and [Pipeline Network Failure = Leakage Failure], then the three candidate 2-item sets obtained are [Pressure Sensor 1 = High, Pressure Sensor 2 = Low], [Pressure Sensor 1 = High, Pipeline Network Failure = Leakage Failure], and [Pressure Sensor 2 = Low, Pipeline Network Failure = Leakage Failure]. Then, calculate the support of each candidate 2-item set according to the compressed transactions, and the 2-item sets with support greater than the preset support are called preferred 2-item sets. Then, repeat the above steps for the preferred 2-item sets to obtain multiple candidate 3-item sets, and judge the support of each candidate 3-item set. If the support of each 3-item set is less than the preset support, that is, there is no preferred 3-item set, then the preferred 2-item sets are used as frequent item sets. If there are preferred 3-item sets, then repeat the above steps for the preferred 3-item sets until no item sets with support greater than the preset support are generated, and multiple frequent item sets are obtained.

[0102] Finally, screen the frequent item sets, and use the frequent item sets containing single-item sets related to pipeline network failures as preliminary association rules, and further evaluate the preliminary association rules according to the preset evaluation indicators to obtain failure association rules.

[0103] Among them, the preset evaluation indicators include confidence, lift, and conviction. When the confidence, lift, and conviction of the preliminary association rule are all greater than the preset values, the preliminary association rule is used as the failure association rule. Confidence is used to measure the probability of the conclusion occurring when the premise condition occurs; lift measures the ratio of the probability of the conclusion occurring when the premise condition appears to the overall probability of the conclusion occurring; conviction is used to measure the degree of difference between the possibility of the conclusion not occurring when the premise condition appears and the overall possibility of the conclusion not occurring.

[0104] Exemplarily, assume that one of the preliminary association rules is [Pressure Sensor 1 = High, Pressure Sensor 3 = Low, Flow Sensor 3 = Low Flow, Pipeline Network Failure = Leakage Failure]. Denote [Pressure Sensor 1 = High, Pressure Sensor 3 = Low, Flow Sensor 3 = Low Flow] as Y and [Pipeline Network Failure = Leakage Failure] as Z. Then the confidence of this preliminary association rule , where represents the support of this preliminary association rule, represents the support of , where represents the support of .

[0105] The method provided in this embodiment generates fault association rules based on the pattern between historical transmission data and historical fault conditions of the pipeline network, and reduces the data processing volume through a transaction compression step to accelerate rule generation. Finally, the quality of the rules is improved through multiple evaluation indicators, and thus the fault types of the underwater pipeline network are accurately obtained.

[0106] Furthermore, this embodiment provides a method for determining the fault types of an underwater intelligent pipeline network according to untransmitted data and fault association rules. First, determine the central detection range of the sensors corresponding to the untransmitted data, obtain the fault association rules corresponding to this range, and obtain the current-time transmission data of other sensors involved in the fault association rules. Discretize the untransmitted data and the current-time transmission data of other sensors, and finally obtain the faults of the underwater intelligent pipeline network within this detection range according to the rules corresponding to the discretization result.

[0107] Furthermore, as Figure 4 shown, the embodiment of the present application provides an underwater intelligent pipeline network fault detection system based on sensor breakpoint detection, including:

[0108] Positioning module: Determine the faulty sensor and its corresponding extension sensors. The faulty sensor is the sensor whose data transmission is interrupted at the current time, and the radiation detection range of the extension sensors includes the central detection range of the faulty sensor;

[0109] Calculation module: Determine the untransmitted data of the faulty sensor at the current time according to the historical transmission data of the faulty sensor and the current-time transmission data of the extension sensors;

[0110] Detection module: Determine the fault association rules according to the historical transmission data of multiple sensors and the historical fault conditions of the underwater intelligent pipeline network, and determine the fault types of the underwater intelligent pipeline network according to the untransmitted data and the fault association rules.

[0111] Among them, the positioning module, the calculation module, and the detection module are all located in the server. The server receives the data transmitted by the acquisition device and conducts further analysis. The acquisition device includes different types of sensors.

[0112] The underwater intelligent pipeline network fault detection method and system provided by the embodiment of the present application accurately restore the untransmitted data of the faulty sensor through the current-time transmission data of the extension sensors and the historical transmission data of the faulty sensor, and accurately identify the fault types of the underwater intelligent pipeline network according to the untransmitted data and the fault association rules.

[0113] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0114] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0116] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of the present invention.

Claims

1. A method for underwater intelligent pipe network fault detection based on sensor breakpoint detection, characterized in that: The underwater smart pipe network is provided with a plurality of sensors, each of which corresponds to a central detection range and a radiation detection range. The underwater smart pipe network fault detection method includes: Determine a faulty sensor and its corresponding extension sensor, wherein the faulty sensor is a sensor whose data transmission is interrupted at a current moment, and a radiation detection range of the extension sensor includes a central detection range of the faulty sensor; Determine the non-transmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment; Determining a fault type of the underwater smart pipe network according to the non-transmitted data and a fault association rule, wherein the fault association rule is determined according to historical transmission data of the multiple sensors and historical fault conditions of the underwater smart pipe network; Wherein, determining the non-transmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment includes: Obtaining first prediction data and second prediction data according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment respectively; Determine the fault type of the fault sensor according to the historical transmission data of the fault sensor and the transmission data of the extension sensor at the current moment, wherein the fault type of the fault sensor includes a sudden fault and a gradual fault; The non-transmitted data is obtained according to the first prediction data, the second prediction data and the fault type of the faulty sensor.

2. According to claim 1, a method for underwater intelligent pipe network fault detection based on sensor breakpoint detection is characterized in that: The obtaining of the first prediction data and the second prediction data according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment respectively comprises: Obtaining the first prediction data according to the historical transmission data of the faulty sensor and a time series model; The second prediction data is obtained according to the transmission data of the epitaxial sensor at the current moment and an interpolation algorithm.

3. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 2 is characterized in that: The second prediction data is obtained according to the transmission data of the epitaxial sensor at the current moment and the interpolation algorithm, including: Determining a triangulated network structure according to the position of the epitaxial sensor; Determine the triangle to which the faulty sensor belongs according to the position of the faulty sensor and the triangulated network structure; Determine a plurality of sub-triangles according to the position of the fault sensor and the triangle; The second prediction data is obtained according to the areas of the multiple sub-triangles and the transmission data of the epitaxial sensor at the current moment.

4. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 1 is characterized in that: The fault association rule is determined according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater smart pipe network, including: Obtaining a transaction data set according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater smart pipe network; determining preliminary association rules based on the transaction data set; The fault association rule is obtained according to the preset evaluation index and the preliminary association rule.

5. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 4 is characterized in that: According to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater smart pipe network, a transaction data set is obtained, including: Determining factors affecting the operation status of the underwater smart pipe network based on the historical transmission data of the multiple sensors; determining a dynamic threshold value based on the factors and historical transmission data of the plurality of sensors; Performing dynamic discretization processing on the historical transmission data of the multiple sensors according to the dynamic threshold to obtain discretized data; According to the historical fault conditions of the underwater intelligent pipe network, the discretized data is organized into the transaction data set.

6. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 4 is characterized in that: The determining of a preliminary association rule based on the transaction data set includes: Based on each item set in the transaction data set, gradually generate multiple candidate item sets; Calculating the support of the multiple candidate item sets, and determining multiple frequent item sets according to the support; A preliminary association rule is determined according to the frequent item sets.

7. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 6 is characterized in that: The step of gradually generating multiple candidate item sets based on each item set in the transaction data set includes: Performing transaction compression processing based on each item set in the transaction data set to obtain multiple compressed transactions; The plurality of candidate item sets are gradually generated based on the plurality of compression transactions.

8. The underwater intelligent pipe network fault detection method based on sensor breakpoint detection according to claim 4 is characterized in that: The preset evaluation indicators include confidence, lift and certainty.

9. An underwater intelligent pipe network fault detection system based on sensor breakpoint detection, characterized in that: The underwater smart pipe network is provided with a plurality of sensors, each of which corresponds to a central detection range and a radiation detection range. The underwater smart pipe network fault detection system includes: Positioning module: determining a faulty sensor and its corresponding extension sensor, wherein the faulty sensor is a sensor whose data transmission is interrupted at the current moment, and the radiation detection range of the extension sensor includes the central detection range of the faulty sensor; A calculation module: determining the non-transmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment; A detection module: determining a fault association rule according to the historical transmission data of the multiple sensors and the historical fault conditions of the underwater smart pipe network, and determining the fault type of the underwater smart pipe network according to the non-transmitted data and the fault association rule; Wherein, determining the non-transmitted data of the faulty sensor at the current moment according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment includes: Obtaining first prediction data and second prediction data according to the historical transmission data of the faulty sensor and the transmission data of the extension sensor at the current moment respectively; Determine the fault type of the fault sensor according to the historical transmission data of the fault sensor and the transmission data of the extension sensor at the current moment, wherein the fault type of the fault sensor includes a sudden fault and a gradual fault; The non-transmitted data is obtained according to the first prediction data, the second prediction data and the fault type of the faulty sensor.

Images