A water plant inspection method and system

Abstract

The application relates to the technical field of water plant inspection, and discloses a water plant inspection method and system. The water quality monitoring module of the method comprises multiple water quality parameter acquisition units and a water quality abnormal fluctuation judgment unit. The multiple acquisition units are arranged on multiple water supply pipelines of the water plant. The abnormal fluctuation judgment unit determines a water quality fluctuation entropy value based on the output of the acquisition units. The equipment state monitoring module is used for judging whether to trigger equipment monitoring based on the water quality fluctuation entropy values of the acquisition units. When triggering, the pressure fluctuation frequency spectrum characteristics of the corresponding pipelines are acquired. The inspection decision module generates a pipeline network risk level and outputs an optimal inspection route based on the pressure fluctuation frequency spectrum characteristics. The system realizes abnormal sensing through water quality parameter acquisition and fluctuation entropy value calculation, analyzes the equipment state in combination with pressure fluctuation frequency spectrum, and then scientifically generates a risk level and an inspection route, so that the accuracy and efficiency of water plant inspection are improved, and the inspection demand of a complex water supply pipeline network is met.

Description

Technical Field

[0001] This invention relates to the field of water plant inspection technology, specifically a water plant inspection method and system. Background Technology

[0002] In the daily operation of a water plant, the stable operation of the water supply network is directly related to the safety and quality of water supply. The water supply network has a complex structure, a wide coverage area, and involves numerous pipes, valves, and monitoring equipment. Its operating status is affected by various factors such as water quality, pressure, and equipment aging, making it prone to various abnormal problems. Traditional water plant inspection methods mostly rely on regular manual inspections. Inspectors carry portable testing equipment to collect parameters at preset points and then judge whether there are any abnormalities in the network based on experience. This method has obvious limitations: on the one hand, the frequency of manual inspections is limited, making it difficult to achieve real-time or high-frequency monitoring of the network, resulting in some sudden anomalies not being detected in time; on the other hand, the setting of inspection points relies on historical experience, which may create monitoring blind spots, making it difficult to detect subtle abnormal fluctuations in local pipelines.

[0003] Water quality parameters are crucial indicators reflecting the operational status of pipe networks, including turbidity, pH, residual chlorine content, and dissolved oxygen. Abnormal fluctuations in these parameters often indicate problems such as pipeline contamination, leaks, or abnormal treatment processes. Traditional water quality monitoring primarily employs fixed-point sampling and analysis, resulting in a limited number and uneven distribution of sampling points, failing to comprehensively reflect the water quality status of the entire pipe network. When slight fluctuations occur in the water quality of a certain area, the lack of continuous monitoring data makes it difficult to quickly identify the source of the anomaly. Furthermore, current technologies often rely on threshold comparisons of single parameters to determine water quality anomalies; that is, an anomaly is identified when a parameter exceeds a preset threshold. However, fluctuations in water quality parameters within actual pipe networks are often correlated and complex, making single-threshold judgments prone to misjudgment or missed detection.

[0004] Equipment status is closely related to water quality. Changes in the operating status of equipment such as pumps and valves in water supply pipelines directly affect pipeline pressure, leading to fluctuations in water quality parameters. Traditional equipment status monitoring and water quality monitoring are independent of each other. Equipment monitoring focuses primarily on its own parameters such as vibration and temperature, without establishing an effective correlation with changes in water quality parameters. When water quality anomalies occur, it is difficult to quickly pinpoint whether the cause is equipment failure, requiring multiple separate tests and checks on both water quality and equipment, increasing the difficulty and time cost of problem diagnosis.

[0005] In terms of inspection route planning, traditional methods often rely on the experience of inspection personnel or fixed routes, without considering the real-time risk distribution of the pipeline network. When multiple potential anomalies exist in the pipeline network, problems such as overlapping inspection routes or omissions of key areas may occur, reducing inspection efficiency. With the expansion of water plant scale and the extension of water supply networks, traditional inspection methods are no longer sufficient to meet the needs of refined and intelligent management. There is an urgent need for an integrated system that can integrate water quality monitoring, equipment status assessment, and inspection decision-making to improve the accuracy of pipeline anomaly identification and the efficiency of inspection work. Summary of the Invention

[0006] The purpose of this invention is to provide a water plant inspection system to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides a water plant inspection system, the method comprising:

[0008] The water quality monitoring module includes multiple water quality parameter acquisition units and a water quality abnormal fluctuation judgment unit. The multiple water quality parameter acquisition units are respectively installed on multiple water supply pipelines of the water plant. The water quality abnormal fluctuation judgment unit is used to determine the water quality fluctuation entropy value of the water quality parameter acquisition unit based on the output of the water quality parameter acquisition unit.

[0009] The equipment status monitoring module is used to determine whether equipment monitoring is triggered based on the water quality fluctuation entropy value of each water quality parameter acquisition unit. When it is determined that equipment monitoring is triggered, the pressure fluctuation spectrum characteristics of the corresponding water supply pipeline are collected.

[0010] The inspection decision module is used to generate the pipeline risk level and output the optimal inspection route based on the pressure fluctuation spectrum characteristics.

[0011] Preferably, the water quality parameter acquisition unit includes a sensor array disposed on the side wall of the water supply pipeline, the sensor array comprising multiple sets of water quality sensing devices, and the detection surface of the sensor array covering the fluid cross-section of the water supply pipeline.

[0012] Preferably, the water quality abnormal fluctuation judgment unit determines the water quality fluctuation entropy value of the water quality parameter acquisition unit based on the output of the water quality parameter acquisition unit, including:

[0013] For each group of water quality sensing devices, the water quality parameter sequence within a continuous time period is extracted, and a reference parameter component and a noise component are generated through a time-domain decomposition algorithm. The reference parameter component is then corrected based on a noise suppression model to generate an optimized water quality parameter sequence.

[0014] Based on the optimized water quality parameter sequence corresponding to all the water quality sensing devices, the water quality fluctuation entropy value of the water quality parameter acquisition unit is calculated.

[0015] Preferably, the equipment status monitoring module determines whether to trigger equipment monitoring based on the water quality fluctuation entropy value, including:

[0016] Calculate the difference in water quality fluctuation entropy values ​​between any two of the water quality parameter acquisition units;

[0017] Based on the discrete distribution characteristics of the differences in entropy values ​​of all water quality fluctuations, water quality anomaly clustering parameters are generated.

[0018] When the water quality anomaly aggregation parameter is less than the preset aggregation threshold, and at least one of the water quality fluctuation entropy values ​​exceeds the preset entropy value threshold, the device monitoring is triggered.

[0019] Preferably, the outer wall of the water supply pipeline is provided with multiple vibration marker points;

[0020] The equipment status monitoring module collects the pressure fluctuation spectrum characteristics, including: acquiring vibration waveform data of the water supply pipeline under multiple pressure detection positions;

[0021] The inspection decision module generates a pipeline risk level based on the pressure fluctuation spectrum characteristics, including:

[0022] For each pressure detection location, analyze the spatial correlation between the pressure peak value and multiple vibration marker points in the vibration waveform data;

[0023] By integrating the spatial relationships corresponding to all pressure detection locations, a comprehensive stress distribution map of the pipeline network is generated.

[0024] The risk level of the pipeline network is determined based on the comprehensive stress distribution map of the pipeline network.

[0025] Preferably, the inspection decision module outputs the optimal inspection route based on the pipeline network risk level, including:

[0026] Multiple candidate inspection routes are generated based on the pipeline network risk level.

[0027] Obtain the distribution density of historical leakage events in the water plant's pipeline network at continuous time points;

[0028] Obtain the water supply load intensity of the water plant at the current time point;

[0029] By combining the distribution density of historical leakage events and the water supply load intensity of the water plant, the multiple candidate inspection routes are screened to generate the optimal inspection route.

[0030] Preferably, it also includes an emergency response module, used to generate inspection priority instructions based on the pipeline network risk level, the distribution density of historical leakage events, and the water supply load intensity of the water plant.

[0031] Preferably, the inspection decision module outputs the optimal inspection route, including:

[0032] Generate preset routine inspection routes based on historical inspection records;

[0033] When the path overlap rate between the optimal inspection route and the preset regular inspection route is lower than the preset overlap threshold, the optimal inspection route and the preset regular inspection route are updated.

[0034] Preferably, the inspection decision module determines the pipeline network risk level based on the comprehensive stress distribution map of the pipeline network, including:

[0035] Extract the area percentage of high-stress regions and the rate of change of stress gradient from the comprehensive stress distribution map of the pipeline network;

[0036] The equipment degradation coefficient is generated by processing the area ratio of the high-stress region and the rate of change of the stress gradient through an equipment degradation assessment algorithm.

[0037] The pipe age parameters and material fatigue index of the water supply pipeline are obtained, and the pipeline network risk level is generated by combining them with the equipment deterioration coefficient.

[0038] Preferably, the present invention also includes a water plant inspection method, which includes all the modules and method flow of the above-mentioned water plant inspection system.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] This water plant inspection system, through the collaborative operation of multiple modules, provides a more intelligent and precise solution for water plant inspection work. In terms of water quality monitoring, the system sets up multiple water quality parameter acquisition units distributed along multiple water supply pipelines, overcoming the limitations of traditional single or limited acquisition points and enabling more comprehensive capture of water quality changes in various areas of the pipeline network. The water quality anomaly fluctuation judgment unit calculates the water quality fluctuation entropy value based on the acquired parameters. Leveraging the quantification capability of entropy values ​​for the complexity of parameter fluctuations, it can more sensitively identify subtle anomalies that are easily overlooked by traditional threshold judgment methods, achieving early detection of water quality anomalies.

[0041] The equipment status monitoring module and the water quality monitoring module work together, not by continuously and indiscriminately monitoring the equipment, but by determining whether to trigger equipment monitoring based on the entropy value of water quality fluctuations. This triggering mechanism makes equipment monitoring more targeted and avoids wasting monitoring resources. When abnormal fluctuations occur in water quality, the pressure fluctuation spectrum characteristics of the corresponding water supply pipeline are collected in a timely manner, linking the water quality anomaly with the pipeline pressure status. This helps to trace the cause of the anomaly from the perspective of equipment operation, reducing the information gap problem when water quality and equipment are detected separately, and making anomaly diagnosis more directional.

[0042] The inspection decision module generates pipeline risk levels based on pressure fluctuation spectrum characteristics, changing the traditional inspection model that relies on experience to judge risk. By analyzing pressure spectrum characteristics, the risk level of different pipelines can be assessed more scientifically, and key inspection areas can be identified. Based on this, the optimal inspection route output takes into account the risk distribution of the pipeline network, reducing route duplication during the inspection process, ensuring that inspection resources are prioritized for high-risk areas, and improving the overall efficiency of the inspection work.

[0043] The entire system integrates water quality monitoring, equipment status assessment, and inspection decision-making, ensuring smooth information flow between modules and forming a complete monitoring-diagnosis-decision closed loop. Compared to traditional inspection methods, this system can detect potential problems in the pipeline network more promptly, reducing the possibility of misjudgments or omissions through multi-parameter linkage analysis. Furthermore, data-driven risk level classification and route planning make inspection work more organized and targeted, adapting to the refined management needs of complex water supply networks and contributing to the stable operation of the water plant's supply network. Attached Figure Description

[0044] Figure 1 This is a timing diagram of the water plant inspection system described in this invention;

[0045] Figure 2 A flowchart for device monitoring trigger judgment;

[0046] Figure 3 A flowchart for collecting pressure fluctuation spectrum characteristics and generating pipeline risk levels;

[0047] Figure 4 A flowchart for updating the inspection route. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 The present invention provides a water plant inspection method and system. The method includes: the water plant inspection system includes a water quality monitoring module, an equipment status monitoring module, and an inspection decision module.

[0050] The water quality monitoring module has multiple water quality parameter acquisition units distributed across various water supply pipelines within the water plant. The water quality anomaly fluctuation judgment unit calculates and determines the water quality fluctuation entropy value based on the output data from each water quality parameter acquisition unit. The equipment status monitoring module analyzes this water quality fluctuation entropy value to determine whether to initiate equipment monitoring; when monitoring is triggered, it acquires the pressure fluctuation spectrum characteristics of the corresponding water supply pipeline. The inspection decision module processes these pressure fluctuation spectrum characteristics, generates a pipeline network risk level, and outputs the optimal inspection route accordingly.

[0051] Example 1: See Figure 2 and Figure 3 The water quality parameter acquisition unit installed on the side wall of the water supply pipeline adopts a ring-shaped sensor array structure. This array contains eight independent water sensor groups, each group evenly arranged at 45-degree circumferential intervals. The sensor types include pH probes, turbidity sensors, residual chlorine detectors, and dissolved oxygen monitors, with two symmetrically distributed groups of each type in the array. The number of sensors increases appropriately with increasing pipe diameter at different locations within the water supply pipeline; the installation depth of the sensors is determined based on the pipe wall thickness to ensure accurate contact with the fluid inside the pipe; a signal isolation module is added between the sensor array and the data transmission line to filter external electromagnetic interference. The sensor detection surface is embedded in the inner wall of the pipeline, covered with a hydrophobic and anti-fouling coating, and the sealing structure uses a fluororubber corrugated pipe. The installation of the sensor array must be adapted to the laying environment of the water supply pipeline. If the pipeline is buried in a damp underground area, the outer shell of the sensor device must be made of 304 stainless steel and treated with anti-corrosion coating. Waterproof tape should be wrapped around the interfaces and waterproof sleeves should be installed to prevent groundwater from seeping in and damaging the internal components. If the pipeline is located in an open area, a sunshade cover must be installed on the outside of the sensor device. The cover should be made of polycarbonate material, which has both sun protection and impact resistance properties, to avoid direct sunlight causing aging of the sensor components or rainwater erosion affecting the detection accuracy. At the same time, the sensor array must be kept at a reasonable distance from the maintenance valve group of the water supply pipeline, usually 1.5-2 meters downstream of the maintenance valve group. This avoids the impact of water flow during valve opening and closing affecting parameter acquisition and facilitates subsequent maintenance and calibration of the sensor device. The detection direction of all sensors is perpendicular to the water flow direction, and the installation axis is parallel to the center line of the water supply pipeline, covering a fluid cross-section with a diameter of 200mm to 1500mm. The sensors are spaced 300mm apart along the axial direction to ensure that the detection area continuously covers more than 70% of the axial cross-section of the pipeline.

[0052] The sensor signal acquisition employs time-division multiplexing technology, completing a full-array data synchronous acquisition cycle every 5 seconds. The data format includes timestamps, sensor IDs, current values, and calibration coefficients. The acquisition process transmits data via RS-485 bus to the water quality anomaly fluctuation judgment unit. This unit has a built-in timing processor that performs independent analysis on each group of water quality sensors. The specific process is as follows: A water quality parameter sequence generated by a single group of sensors over a continuous 72-hour period is extracted. This sequence contains 51,840 data points (sampling interval of 5 seconds). The timing processor uses a time-domain decomposition algorithm to process the sequence. The core of this algorithm includes a three-stage recursive filtering structure: the first stage uses a low-pass filter with a 0.1Hz cutoff frequency to extract the baseline parameter component, which characterizes the long-term trend of water quality parameters; the second stage uses a high-pass filter to separate the noise component from the remaining signal, with the noise component frequency distributed in the 0.5Hz-10Hz range; the third stage reconstruction system generates a complementary signal sequence based on the first and second stage outputs to eliminate baseline drift. The noise suppression model operates on the baseline parameter components. The model incorporates an adaptive threshold calculator that analyzes the root mean square (RMS) value of the noise components in real time and dynamically adjusts the correction coefficients. When the RMS of the noise component exceeds a preset noise threshold, a Gaussian smoothing algorithm is activated to process the baseline parameter components; when the noise level is below the threshold, the original baseline parameter component data characteristics are preserved. Noise suppression sensitivity is increased for turbidity parameters, and noise suppression intensity is decreased for pH parameters, while preserving the inherent variation characteristics of the parameters themselves. The corrected baseline parameter components and the processed noise components are superimposed to generate an optimized water quality parameter sequence, with the sequence time resolution remaining constant at 5 seconds.

[0053] The calculation of water quality fluctuation entropy is based on the optimized sequence data of all eight sets of devices. The system creates a 256-second sliding time window (containing 51 sampling points) and simultaneously extracts data segments of the eight sets of sequences within the window. After normalizing each set of sequence data to the [0,1] interval, it is input into the entropy calculator, which performs discrete probability distribution analysis: first, the data range of each set is divided into 50 equal intervals, and the frequency of data points in each interval is counted; second, a probability distribution function is constructed using the Shannon entropy formula, and the entropy value of each of the eight sets of sequences within the time window is calculated; finally, the arithmetic mean of the eight sets of entropy values ​​is used as the instantaneous entropy value output of the water quality parameter acquisition unit. This calculation process updates the output results every 10 seconds, forming a continuous fluctuation entropy curve for the equipment status monitoring module to call. During the operation of the sensor array, the in-situ calibration procedure is periodically initiated. During the calibration process, each set of sensors is measured synchronously with the standard reference probe, and the deviation correction coefficient is dynamically updated to the calibration database. Before in-situ calibration, the system automatically closes the flow regulating valves of the corresponding pipelines, reducing the water flow velocity to below 0.5 m / s to ensure stable water quality parameters during calibration. The standard reference probe used for calibration must be calibrated in the laboratory beforehand, with the calibration error controlled within ±2%. After each calibration, the system generates a calibration report, recording the calibration time, the sensor numbers involved, the deviation value, and the correction factor. This report is automatically archived in the calibration archive of the water quality monitoring module for easy traceability of calibration records. Furthermore, when a sensor has been running continuously for more than 6 months, it must be manually removed for offline calibration. During offline calibration, the sensor is placed in standard solutions of different concentration gradients to verify its accuracy in detecting parameters such as turbidity, pH, and residual chlorine. If the detection error of any parameter exceeds 5%, the sensor needs to be adjusted or replaced. Data storage employs a dual-buffering mechanism, storing currently acquired data and historical optimized sequences in separate databases to ensure uninterrupted continuous computation.

[0054] Example 2: The equipment status monitoring module receives water quality fluctuation entropy data streams from the water quality monitoring module. This data is continuously input in time-series format with a sampling interval of 10 seconds. The input data structure includes three parts: identifier, timestamp, and entropy measurement result, covering water quality parameter acquisition units distributed across 32 water supply pipelines within the plant area. During the module initialization phase, a data buffer is established with a capacity designed to accommodate all entropy records within a 2-hour continuous sampling period. The data buffer adopts a dual-partition storage structure, divided into a real-time data area and a backup data area. The real-time data area stores the currently processed entropy data, while the backup data area synchronously backs up the contents of the real-time data area. If the real-time data area loses data due to a sudden power outage or system failure, the backup data area can immediately switch to the real-time data area to continue working, ensuring data continuity. Simultaneously, the data buffer needs to be defragmented periodically, with a defragmentation program initiated every 24 hours to clear redundant data and erroneous records, freeing up storage space and preventing data overflow due to insufficient buffer space. The processing flow begins with a data format verification program, which automatically filters invalid identifier records and corrects timestamp misalignments.

[0055] The calculation of water quality fluctuation entropy difference is performed within a fixed time window. The time window is set to a 5-minute interval, and the latest 30 entropy measurement results of each acquisition unit within that time period are automatically acquired at the start of each calculation round. The object of difference calculation is defined as all unit pairs in the water supply network. For a system with N acquisition units, C(N,2) groups of unit comparisons need to be processed. The entropy value changes of the same acquisition unit within adjacent time windows are compared. If there are large fluctuations in a short period of time, the unit is marked as a potential abnormal unit. The operation for each group of unit comparisons is as follows: extract the entropy value sequence of unit A and unit B within the current time window, calculate the absolute difference of the entropy value at each corresponding time point in the sequence, perform linear mean calculation on all 30 differences, and record the output result as the original difference value. After all unit pairs have completed this operation, a difference matrix is ​​formed. This matrix is ​​stored in a lower triangular compressed structure to save memory resources. The update frequency of the difference matrix is ​​synchronized with the time window movement frequency, that is, a complete matrix is ​​regenerated every 5 minutes.

[0056] The water quality anomaly clustering parameter generation mechanism is based on the construction of a difference matrix. First, the set of all original difference values ​​in the matrix is ​​extracted, and this data set is standardized: the maximum-minimum method is used to normalize all original values ​​to the [0,1] interval. The parameter calculation algorithm uses spatial clustering technology to construct a three-dimensional feature vector: dimension one is the mean of the difference distribution, calculated as the arithmetic mean of the standardized data; dimension two is the distribution dispersion, output using the standard deviation formula; dimension three is the distribution skewness, analyzed using the third-order central moment formula to assess data asymmetry. After feature vector construction, it is input into the clustering analysis engine, which uses a density clustering algorithm to automatically identify anomalous feature clusters. Adjacent acquisition units are divided into regional groups according to the pipeline geographical layout, and water quality anomaly clustering parameters are calculated for each group. Algorithm parameter settings include preset values ​​such as a neighborhood search radius of 0.2 and a minimum number of neighbors for the core point of 12. The output results are the number of detected anomalous clusters and the cluster density index. The water quality anomaly clustering parameter is ultimately defined as a weighted product of the cluster density index and the number of clusters, with the weighting coefficient set as a fixed proportion according to the water plant operation procedures.

[0057] The equipment monitoring triggering mechanism operates under a dual judgment framework of abnormal water quality aggregation parameters and entropy values. The system maintains a database of preset aggregation thresholds, with a fixed benchmark of 0.85 based on the pipeline network scale. Simultaneously, it monitors the water quality fluctuation entropy levels of 32 data acquisition units in real time, setting a preset entropy threshold of 0.7 as the standard limit. The judgment logic includes two parallel detection branches: branch one continuously compares the relationship between the abnormal water quality aggregation parameters and the preset aggregation threshold; branch two scans whether the current entropy value of each unit exceeds the preset entropy threshold. The system performs a full-condition check every 10 seconds: when the latest value of the abnormal water quality aggregation parameter is below 0.85, and at least one data acquisition unit has a real-time entropy value exceeding 0.7, the equipment monitoring operation is triggered. The preset entropy threshold is increased during the rainy season and decreased during the dry season to adapt to the water quality characteristics of different seasons. After the trigger signal is generated, it is automatically bound to the water supply pipeline number that meets the entropy value exceeding condition, establishing a monitoring task instruction queue. The monitoring task command queue adopts a priority sorting mechanism. If multiple water supply pipelines trigger equipment monitoring simultaneously, the system will sort them according to the water supply priority of the pipelines. The water supply priority is determined by the water supply range of the pipelines. Pipelines supplying water to residential areas have a higher priority than pipelines supplying water to industrial auxiliary water areas, and pipelines supplying water to key locations such as hospitals and schools have the highest priority. Before sending a task command, the system will first check whether the pressure monitoring equipment of the corresponding water supply pipeline is online. If the equipment is offline, it will immediately send an offline alarm message to the mobile terminal of the maintenance personnel and attempt to reconnect to the equipment. If the connection fails three times in a row, the monitoring task of that pipeline will be marked as "pending" and the task queue order will be adjusted to prioritize the monitoring tasks of pipelines with online equipment. The pressure fluctuation spectrum acquisition equipment starts immediately after receiving the enqueue command, accurately locating the acquisition position to a specific monitoring section of the target pipeline, and the task execution status is transmitted back to the operation log in real time. In the non-triggered state, the system enters a low-power standby mode, retaining a port for dynamic adjustment of preset thresholds to support remote parameter configuration. This monitoring triggering mechanism controls the running cycle through an independent timer, keeping it decoupled from the upstream data processing sequence to ensure the immediacy of the judgment and response.

[0058] Example 3, see Figure 4The process of generating the optimal inspection route in the inspection decision module begins with receiving the pipeline risk level data structure, which is divided into four levels. The module loads the digital pipeline topology map of the water plant as the underlying model. This model is stored in the spatial geographic information system database, recording the spatial coordinates and attribute information of all water supply pipeline connection nodes. Based on the received pipeline risk level labeling results, the system labels high-risk pipelines to form key areas of concern. The coverage area is automatically generated by combining pipeline segments with risk levels of three or above. The candidate inspection route generation engine uses a heuristic path planning algorithm to construct a set of routes. The specific method is as follows: starting from the coordinate origin of the water plant control center, the main and branch pipelines in different directions are scanned sequentially; for each main and branch pipeline, the path segment is divided with a preset 3-kilometer basic detection unit; all high-risk area path segments are traversed through depth-first search to form an initial path set; path segments from adjacent non-high-risk areas are added to expand the route diversity; finally, a set of K candidate routes is generated, with K fixed at 8 and covering branch routes in different directions.

[0059] Historical leakage event distribution density data is acquired via an event log analyzer. The system connects to the water plant safety monitoring database interface to extract a dataset of pipeline leakage events recorded within the last 90 days. This data includes structured fields such as event occurrence time, geographical coordinates, and leakage level. Data processing includes spatial gridding: dividing the plant's planar grid into a 1-meter precision cell array; calculating time-dimensional sliding window statistics; analyzing the number of events within each grid cell using a 72-hour cycle; and generating an event density heatmap layer stored in a spatial data warehouse. The real-time water plant supply load intensity acquisition module obtains the main pipeline flow velocity value through a flow sensor network and calculates and outputs the load intensity index parameter, expressed by the formula:

[0060]

[0061] in: This represents the load intensity index at time t. The total number of monitoring points. It is the real-time flow velocity value at the i-th monitoring point. Represents the maximum safe flow rate at design speed. The weighting coefficient is set based on the importance of the area where the monitoring point is located.

[0062] The route selection and optimization model operates a multi-objective decision-making mechanism on a set of candidate routes. The objective function integrates three key parameters: the first parameter... The total length of high-risk road segments in candidate route k is quantified; the second parameter. Mapping the historical leakage event distribution density integral value of the area covered by this route; third parameter This reflects the average load intensity index of the area corresponding to the route. The optimization objective is:

[0063]

[0064] Where: coefficient set The pipeline weight configuration table is preset. During execution, all candidate routes are sorted by objective function value calculation, and the route with the smallest value is selected as the initial optimal route. This route is submitted to the dynamic verification stage for matching and verification with the current pipeline operation status: real-time operation information of the water plant scheduling is obtained to avoid pipeline sections where valve operations are in progress; the dynamic verification stage also needs to access the operation status data of auxiliary equipment of the water supply pipeline, including the operating power of water pumps, the on / off status of valves, and the working status of flow meters. If the water pump of a certain pipeline section is under maintenance, the water pressure and flow velocity of that pipeline section will fluctuate abnormally. At this time, the pipeline section needs to be temporarily removed from the inspection route. After the water pump is repaired and resumes normal operation, it will be included in the inspection scope again; if the flow meter shows that the flow of a certain pipeline section is zero, it means that the pipeline section is in a shutdown state and does not need to be inspected, thus avoiding the waste of inspection resources. Meanwhile, dynamic verification also needs to consider the working time constraints of inspection personnel. If the estimated inspection time of a candidate inspection route exceeds the single working time of an inspection personnel (usually set to 4 hours), the system will automatically split the route into two or more segments and arrange inspections in different time periods. The estimated inspection time of each segment is controlled within 3.5 hours, with 0.5 hours reserved for emergency handling. The system connects to the traffic control system to identify road traffic status; the final output is a path sequence containing three-dimensional spatial coordinates, with the total length of the sequence controlled within the specified range. The path sequence includes a time planning table, which divides the system into 6 standard inspection time periods, each with a list of pipeline inspection tasks to be completed. The schedule is dynamically adjusted based on the season and daily peak water supply periods. For example, during peak summer water usage periods (typically 8:00-10:00 and 18:00-20:00), the pressure and flow rate of the water supply pipelines fluctuate frequently. Therefore, inspection times are adjusted to avoid peak usage periods, such as 2:00-6:00 AM or 10:00-14:00 AM, minimizing the impact of inspections on the normal water supply network. In winter, due to lower temperatures, water supply pipelines may be at risk of freezing. The schedule increases the frequency of inspections of exposed pipelines and pipelines at the network's end points, scheduling these inspections during the warmer hours of 12:00-15:00 to facilitate inspections for freezing cracks and icing. Furthermore, the schedule includes a 15-minute buffer for each inspection period. If the previous inspection task is completed ahead of schedule, the buffer time can be used to begin the next task; if the previous task is delayed, it can be made up within the buffer time, ensuring the overall inspection plan proceeds smoothly. The output results are transmitted to the mobile inspection terminal device via a message queue, and simultaneously backed up to the inspection history archive for subsequent analysis. The system synchronously updates route selection records and correction coefficient sets. The weighted factor configuration and adaptive optimization algorithm perform weight fine-tuning every 24 hours to maintain the continuous adaptability of route generation.

[0065] Example 4: The water plant emergency response module operates in a water supply network system, which includes seven main water supply pipelines, labeled A01 to A07. Input parameters include the current network risk level mapping table, historical leakage event spatial distribution dataset, and real-time water supply load intensity parameters. The data acquisition cycle is set to 15 minutes, and the system time is calibrated to 11:00 AM Beijing time on August 15, 2025. For specific parameters, please refer to Table 1.

[0066] Table 1: The parameters of the seven main water supply pipelines are as follows.

[0067]

[0068] The emergency response module executes priority command generation procedures as follows: Based on the risk level conversion coefficient table, level 4 risk is converted to a 9-point value, level 3 to a 7-point value, level 2 to a 5-point value, and level 1 to a 3-point value. Leakage density index and load intensity are directly input using a percentage system. The calculation model assigns a weight of 60% to risk level, 25% to leakage density, and 15% to load intensity. Example calculation for pipeline A01: Risk value 9 × 0.6 = 5.4, leakage value 0.85 × 25 = 21.25, load value 0.92 × 15 = 13.8, totaling 40.45 urgency score. After the scores for the seven pipelines are calculated, a score ranking queue is formed: A06 (43.5 points) > A01 (40.45 points) > A07 (38.45 points) > A04 (37.25 points) > A05 (35.5 points) > A02 (33.9 points) > A03 (27.45 points). The module generates an inspection priority instruction sequence: A06→A01→A07→A04→A05→A02→A03. When generating candidate routes, it avoids construction areas and traffic congestion sections, selecting routes with good road conditions. The instructions are pushed to the mobile inspection terminal via an encrypted data link.

[0069] Upon receiving the emergency response command, the inspection decision module activates the route generation engine. The system searches the preset regular inspection route database and finds that the standard route for the water plant is defined as: Central Control Room → A01 → A03 → A05 → A02 → Equipment Warehouse. The optimal inspection route generated based on the priority command is: Central Control Room → A06 → A01 → A07 → A04 → Equipment Warehouse. The path overlap rate is calculated using a graph theory comparison algorithm: first, the route is decomposed into a node sequence. The regular route sequence contains five nodes, C1 to C5; the optimal route contains five nodes, O1 to O5. Overlapping nodes must meet two conditions: the position coordinate deviation is less than 2 meters and the node identifiers match completely. Only the Equipment Warehouse node meets the conditions, resulting in one overlapping node. The overlap rate calculation formula is: Number of overlapping nodes / min (total number of route nodes) = 1 / min(5,5) = 20%. The system's preset overlap rate threshold is 40%. The current overlap rate of 20% is below the threshold, triggering a route update procedure.

[0070] The route update operation includes a dual-track parallel mechanism: the main thread saves the current optimal inspection route as a new preset regular inspection route, overwriting the original database record. A log is established to record the reason, time, pipelines involved, and route comparison information for each route update. The new route is structured as a node sequence array: [starting point coordinates (115.2, 38.5), A06 well coordinates (114.8, 39.1), A01 valve coordinates (115.0, 38.7), A07 monitoring station coordinates (115.3, 39.0), A04 hub point coordinates (114.9, 39.2), equipment warehouse coordinates (115.1, 38.9)]. Each coordinate point in the node sequence array is accompanied by detailed on-site identification information, including the pipeline specifications (diameter, wall thickness), installation time, historical maintenance records, and surrounding environment description. For example, the identification information for the A06 well coordinate point will indicate that the pipeline corresponding to the well has a diameter of DN800, a wall thickness of 12mm, was installed in 2018, underwent internal anti-corrosion treatment in 2022, and that there is a municipal sewage well within 50 meters, requiring close monitoring for potential sewage infiltration risks. This identification information is simultaneously pushed to the mobile terminals of inspection personnel. Upon arrival at the site, inspection personnel can view detailed information through their terminals to quickly understand the pipeline situation at that location, improving the targeting of inspections. Simultaneously, a timestamp "202508151130" is added to the route index table to mark the update time. An auxiliary thread dynamically adjusts the optimal inspection route: it accesses the weather forecast interface to obtain the rainfall intensity forecast for the next 3 hours, automatically inserting flood control checkpoints when rainfall exceeds 50mm / h; it accesses water pollution monitoring data around the pipeline network, adjusting routes and increasing inspection frequency when pollution risks are detected. The pipeline pressure monitoring system updates the risk hotspot distribution in real time, expanding inspection routes when new risk points of level three or above are identified. Updated route data packets are written to a distributed storage system using binary encoding, with historical version archives retained for 30 days. Historical version archives are compressed, compressing the node sequence, generation time, corresponding risk level data, and selection criteria for each historical inspection route into ZIP files with a compression ratio of approximately 1:5 to save storage space. Archived files are named "Inspection Route_Update Date_Version Number.zip", for example, "Inspection Route_20250815_V1.0.zip", facilitating quick retrieval of historical routes for specific dates. The system also periodically checks the integrity of archived files, initiating a check every 7 days. If a file is found to be corrupted or missing, it automatically retrieves the corresponding backup file from the backup server for repair, ensuring the integrity of historical version data. The system synchronously pushes route change notifications to the management terminal, including a heatmap comparing the old and new routes and a curve showing the overlap rate. Detailed explanations of the risk levels for each pipeline are attached to the route change notification.

[0071] Example 5: The input data for the comprehensive stress distribution map of the pipeline network comes from the dynamic analysis results generated by the equipment condition monitoring module. This map is stored in a three-dimensional spatial matrix structure with a unit resolution of 5 cm. The construction of the three-dimensional spatial matrix requires the integration of three-dimensional modeling data of the water supply pipeline. The modeling data includes the spatial coordinates, direction, turning angles, and connection relationships with other pipelines of the pipeline, ensuring that each unit in the matrix accurately corresponds to the specific location of the actual pipeline. The update frequency of the matrix data is consistent with the acquisition frequency of the pressure fluctuation spectrum characteristics, updating once every 10 minutes to ensure that the comprehensive stress distribution map of the pipeline network can reflect the stress changes of the pipeline in real time. In addition, the three-dimensional spatial matrix also supports layered display, which can display the axial stress, radial stress, and circumferential stress distribution of the pipeline separately, facilitating the inspection decision module to analyze the stress state of the pipeline in different directions in more detail. After loading the matrix, the data processing engine performs a high-stress area identification operation: using a pixel classification method to identify all area units with stress values ​​exceeding 35 MPa, calculating the proportion of these units in the total matrix and converting it into an area ratio value. An image segmentation algorithm is used to accurately divide the boundary between high-stress areas and normal areas, improving the accuracy of area ratio calculation. Simultaneously, the process of calculating the stress gradient change rate parameter involves spatial vector field analysis: a detection baseline is set along the pipeline axis, and sampling points are taken at 10-centimeter intervals; a ray path of 15 centimeters in length is established along the normal direction at each point; the stress value change process is measured along the ray, and the average stress change per centimeter is calculated, which serves as the raw output value of the gradient change rate. Stress gradient measurements are added along the pipeline circumference to comprehensively obtain the stress change situation.

[0072] The equipment degradation assessment algorithm operates using a three-level fuzzy logic inference framework. The first-level system input consists of two parameters: the area ratio of high-stress regions and the stress gradient change rate. Membership functions are established for each parameter: the area ratio function has three fuzzy sets (low, medium, and high), with boundary points set at 15% and 25%; the gradient change rate function has three levels (stable, fluctuating, and severe), with boundaries of 0.3 MPa / cm and 0.6 MPa / cm. After fuzzification, the input parameters activate 128 inference rules in the rule base. The rules are synthesized using the MAX-MIN method to generate the output fuzzy set. The second-level defuzzification process uses the centroid method to calculate the accurate degradation coefficient output, which is normalized to a continuous numerical range from 0 to 10.

[0073] Pipe age parameters are obtained in real time through the water plant asset management system interface. Data items include the pipe section's manufacturing date and installation time. The total service life in months is calculated by subtracting the installation date from the current date. Combined with pipeline maintenance records, the effective pipe age is recalculated based on the repair time for partially repaired or replaced pipes. Material fatigue index calculation is based on matching with the pipeline material database: material characteristic records are indexed according to the pipeline number, and standard fatigue life curve parameters are extracted; combined with historical data on the number of pressure cycles the pipeline has endured during its service life, the current damage factor is calculated according to the Miner linear damage criterion; this factor is compared with the material fatigue limit and mapped to a fatigue index in the range of 0.0 to 1.0. Determining the material fatigue limit requires reference to the pipeline material's factory inspection report, which must include the material's fatigue life curves under different temperature and pressure conditions. The system will extract the corresponding fatigue limit value from the curves based on the actual operating temperature (typically 5-25℃) and working pressure (typically 0.3-0.6MPa) of the water plant's supply pipeline. If the pipeline experiences overpressure operation (pressure exceeding the design pressure by 1.2 times) during operation, the system will add an overpressure influence coefficient when calculating the fatigue index. The longer the overpressure operation time and the greater the overpressure magnitude, the larger the influence coefficient, thus reflecting the additional impact of overpressure on pipeline material fatigue. Simultaneously, the material fatigue index is correlated with the pipeline's service life; for every 5 years of service life, the fatigue index increases by 0.1, ensuring that the fatigue index accurately reflects the material deterioration after long-term pipeline use.

[0074] The pipeline risk level classifier is built based on a multivariate regression model. The model input vector contains three parameters: equipment deterioration coefficient, pipe age parameter (converted to years), and material fatigue index. During data preprocessing, standardization is performed to ensure all parameter values ​​are within the same dimension range. The model structure employs a three-layer feedforward neural network: the input layer contains three nodes corresponding to the three parameters; the hidden layer has eight nodes using the sigmoid activation function; and the output layer has four nodes corresponding to the four-level risk level probability distribution. The training dataset covers 1024 historical samples, and the network optimizes the weight parameters using a backpropagation algorithm. During system runtime, the current input parameters are input into the trained network model, and the index corresponding to the maximum value among the four nodes in the output layer is used to determine the current pipeline risk level. Data on pipeline daily maintenance frequency and maintenance effectiveness evaluation are incorporated to adjust the risk level determination results. A confidence index is added to the output data, calculated based on the entropy value of the output probability distribution. If the confidence level is below 85%, a secondary manual verification process is initiated. The risk determination results are converted into structured data packets and transmitted via a message bus to the inspection decision system for collaborative work with the emergency response center system.

[0075] Example 6: In a large water plant water supply network system, the system covers 12 main water supply pipelines and 36 branch pipelines. The water plant inspection system used in this case includes a water quality monitoring module, an equipment status monitoring module, an inspection decision module and an emergency response module. Each module works together in accordance with the preset process to achieve a comprehensive inspection of the pipeline network.

[0076] The water quality monitoring module's water quality parameter acquisition units are installed on 12 main water supply pipelines and 16 key branch pipelines of the water plant, according to design requirements. Each acquisition unit employs a sensor array structure, consisting of pH, turbidity, residual chlorine, and dissolved oxygen sensors. Each sensor array contains six sets of water quality sensors, evenly distributed on the sidewalls of the water supply pipelines. Their detection surfaces completely cover the fluid cross-section of the pipelines, enabling comprehensive capture of water quality parameters at different locations within the pipelines. During daily operation, each water quality sensor continuously collects water quality parameters at its corresponding location at fixed time intervals, and the collected parameter data is transmitted in real time to the water quality anomaly fluctuation judgment unit.

[0077] After receiving parameter data from each water quality sensor, the water quality anomaly fluctuation judgment unit first extracts the water quality parameter sequence collected by each sensor over a continuous 24-hour period. Then, it processes this parameter sequence using a time-domain decomposition algorithm, decomposing it into a baseline parameter component and a noise component. The baseline parameter component reflects the basic trend of water quality parameters during this time period, while the noise component mainly contains parameter fluctuations caused by various interference factors. To improve the accuracy of subsequent calculations, the baseline parameter component is corrected based on a noise suppression model. Through feature analysis of the noise component, the noise-affected portion of the baseline parameter component is removed, thereby generating an optimized water quality parameter sequence. After obtaining the optimized water quality parameter sequences for all water quality sensors, the water quality fluctuation entropy value for each water quality parameter acquisition unit is calculated according to preset calculation rules, integrating all sequence data. This entropy value effectively reflects the fluctuation of water quality parameters over a period of time.

[0078] The equipment status monitoring module receives water quality fluctuation entropy values ​​from each water quality parameter acquisition unit of the water quality monitoring module in real time, and uses this as a basis to determine whether to trigger equipment monitoring. Specifically, it first calculates the difference in water quality fluctuation entropy values ​​between any two water quality parameter acquisition units. By comparing the differences in entropy values ​​between different acquisition units, it gains a preliminary understanding of the distribution of water quality fluctuations within the pipeline network. Next, based on the discrete distribution characteristics of all water quality fluctuation entropy value differences, it analyzes the overall distribution pattern of these differences, thereby generating a water quality anomaly aggregation parameter. This parameter reflects the degree of aggregation of water quality anomalies within the pipeline network. When the water quality anomaly aggregation parameter is less than a preset aggregation threshold, and at least one water quality fluctuation entropy value exceeds the preset entropy threshold simultaneously, the equipment status monitoring module determines that equipment monitoring needs to be triggered.

[0079] After triggering equipment monitoring, the equipment status monitoring module begins collecting the pressure fluctuation spectrum characteristics of the corresponding water supply pipeline. Previously, multiple vibration markers were placed at certain intervals on the outer wall of the water supply pipeline; these markers provide location references for subsequent collection and analysis of pressure fluctuation spectrum characteristics. The equipment status monitoring module acquires vibration waveform data of the water supply pipeline from multiple pressure detection positions using dedicated detection equipment. Each detection position corresponds to a different detection angle on the pipeline to ensure comprehensive capture of pipeline vibration, thereby obtaining complete pressure fluctuation spectrum characteristic data.

[0080] After receiving the pressure fluctuation spectrum characteristic data transmitted by the equipment condition monitoring module, the inspection decision module begins to generate the pipeline network risk level and plan the optimal inspection route. For the vibration waveform data acquired at each pressure detection location, the inspection decision module deeply analyzes the spatial correlation between the pressure peak and multiple vibration marker points. By analyzing the relative positional relationship between the pressure peak and the vibration marker points, it understands the pressure conditions experienced by different parts of the pipeline. Subsequently, it merges the spatial correlations corresponding to all pressure detection locations, comprehensively considers the pressure distribution characteristics of the pipeline under various detection angles, and generates a comprehensive stress distribution map of the pipeline network. This map can intuitively display the stress magnitude experienced by different areas within the pipeline network.

[0081] Based on the comprehensive stress distribution map of the pipeline network, the inspection decision module further extracts the area proportion of high-stress regions and the stress gradient change rate, which are important parameters for assessing the condition of pipeline equipment. The high-stress area proportion and stress gradient change rate are processed using an equipment deterioration assessment algorithm. Combined with the algorithm's internal preset assessment model and rules, an equipment deterioration coefficient is generated, reflecting the degree of deterioration of the pipeline equipment. Simultaneously, the inspection decision module obtains the pipe age parameters and material fatigue index of the water supply pipelines from the water plant's equipment management database. The pipe age parameter reflects the length of time the pipeline has been in use, while the material fatigue index reflects the degree of fatigue experienced by the pipeline material during long-term use. By combining the equipment deterioration coefficient, pipe age parameters, and material fatigue index, and according to preset risk level classification standards, a pipeline risk level is generated, clarifying the risk level of different areas of the pipeline network.

[0082] After generating the pipeline network risk level, the inspection decision module initially generates multiple candidate inspection routes based on this risk level. Each candidate route covers different areas within the pipeline network, with a focus on covering high-risk areas. To further optimize the inspection routes, the module obtains historical leakage event distribution density data of the pipeline network at continuous time points from the water plant's historical operation and maintenance database. By analyzing the distribution of historical leakage events in different time periods and regions, it understands the distribution characteristics of areas with high leakage risk within the pipeline network. Simultaneously, it acquires the water plant's water supply load intensity in real time. The water supply load intensity reflects the current load on the pipeline network during water supply; different load intensities will have different impacts on the network's operating status.

[0083] Based on the distribution density of historical leakage events and the current water supply load intensity of the water plant, the inspection decision module filters the multiple candidate inspection routes initially generated. During the screening process, routes that can cover areas with a high incidence of historical leakage events and can adapt to the current water supply load intensity are given priority. At the same time, the rationality and efficiency of the inspection route are taken into account to avoid the occurrence of duplicate routes or omission of key areas, and finally the optimal inspection route is generated.

[0084] The emergency response module generates inspection priority instructions based on the generated pipeline risk level, the distribution density of historical leakage events, and the current water plant supply load. These instructions clarify the order of inspection work in different areas, enabling inspection personnel to prioritize areas with higher risks, greater potential for leakage, and those significantly affected by the current water supply load, ensuring more targeted inspections.

[0085] The inspection decision module also generates preset routine inspection routes based on historical inspection records. It compares the generated optimal inspection route with the preset routine inspection routes and calculates the path overlap rate. When the path overlap rate is lower than a preset overlap threshold, the inspection decision module automatically updates the optimal inspection route and stores the updated optimal route as a new preset routine inspection route in the system database for reference in subsequent inspections. This ensures that the inspection routes can be continuously optimized as the pipeline network operation changes, maintaining high applicability and effectiveness.

[0086] Throughout the inspection process, the various modules maintained efficient data transmission and collaborative work. The water quality monitoring module continuously provided accurate water quality data to subsequent modules, the equipment status monitoring module triggered equipment monitoring and collected pressure fluctuation data in a timely manner based on the water quality data, the inspection decision module generated a reasonable risk level and inspection route by integrating multiple data sources, and the emergency response module generated inspection priority instructions based on relevant data. Together, they ensured the accuracy and efficiency of the water plant's water supply network inspection work and ensured the stable and safe operation of the network.

[0087] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0088] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A water plant inspection system, characterized by, include: The water quality monitoring module includes multiple water quality parameter acquisition units and a water quality abnormal fluctuation judgment unit. The multiple water quality parameter acquisition units are respectively installed on multiple water supply pipelines of the water plant. The water quality abnormal fluctuation judgment unit is used to determine the water quality fluctuation entropy value of the water quality parameter acquisition unit based on the output of the water quality parameter acquisition unit. The equipment status monitoring module is used to determine whether equipment monitoring is triggered based on the water quality fluctuation entropy value of each water quality parameter acquisition unit. When it is determined that equipment monitoring is triggered, the pressure fluctuation spectrum characteristics of the corresponding water supply pipeline are collected. The inspection decision module is used to generate the pipeline risk level and output the optimal inspection route based on the pressure fluctuation spectrum characteristics. The water quality parameter acquisition unit includes a sensor array disposed on the side wall of the water supply pipeline. The sensor array contains multiple sets of water quality sensing devices, and the detection surface of the sensor array covers the fluid cross-section of the water supply pipeline. The water quality abnormal fluctuation judgment unit determines the water quality fluctuation entropy value of the water quality parameter acquisition unit based on the output of the water quality parameter acquisition unit, including: For each group of water quality sensing devices, the water quality parameter sequence within a continuous time period is extracted, and a reference parameter component and a noise component are generated through a time-domain decomposition algorithm. The reference parameter component is then corrected based on a noise suppression model to generate an optimized water quality parameter sequence. Based on the optimized water quality parameter sequence corresponding to all the water quality sensing devices, the water quality fluctuation entropy value of the water quality parameter acquisition unit is calculated. The equipment status monitoring module determines whether to trigger equipment monitoring based on the water quality fluctuation entropy value, including: Calculate the difference in water quality fluctuation entropy values ​​between any two of the water quality parameter acquisition units; Based on the discrete distribution characteristics of the differences in entropy values ​​of all water quality fluctuations, water quality anomaly clustering parameters are generated. When the water quality anomaly aggregation parameter is less than the preset aggregation threshold, and at least one of the water quality fluctuation entropy values ​​exceeds the preset entropy value threshold, the device monitoring is triggered. The outer wall of the water supply pipeline is equipped with multiple vibration markers. The equipment status monitoring module collects the pressure fluctuation spectrum characteristics, including: acquiring vibration waveform data of the water supply pipeline under multiple pressure detection positions; The inspection decision module generates a pipeline risk level based on the pressure fluctuation spectrum characteristics, including: For each pressure detection location, analyze the spatial correlation between the pressure peak value and multiple vibration marker points in the vibration waveform data; By integrating the spatial relationships corresponding to all pressure detection locations, a comprehensive stress distribution map of the pipeline network is generated. The risk level of the pipeline network is determined based on the comprehensive stress distribution map of the pipeline network.

2. The water plant inspection system of claim 1, wherein, The inspection decision module outputs the optimal inspection route based on the pipeline network risk level, including: Multiple candidate inspection routes are generated based on the pipeline network risk level. Obtain the distribution density of historical leakage events in the water plant's pipeline network at continuous time points; Obtain the water supply load intensity of the water plant at the current time point; By combining the distribution density of historical leakage events and the water supply load intensity of the water plant, the multiple candidate inspection routes are screened to generate the optimal inspection route.

3. The water plant inspection system according to claim 2, characterized in that, It also includes an emergency response module, which is used to generate inspection priority instructions based on the risk level of the pipeline network, the distribution density of the historical leakage events, and the water supply load intensity of the water plant.

4. The water plant inspection system of claim 3, wherein, The inspection decision module outputs the optimal inspection route, including: Generate preset routine inspection routes based on historical inspection records; When the path overlap rate between the optimal inspection route and the preset regular inspection route is lower than the preset overlap threshold, the optimal inspection route and the preset regular inspection route are updated.

5. The water plant inspection system of claim 4, wherein, The inspection decision module determines the pipeline network risk level based on the comprehensive stress distribution map of the pipeline network, including: Extract the area percentage of high-stress regions and the rate of change of stress gradient from the comprehensive stress distribution map of the pipeline network; The equipment degradation coefficient is generated by processing the area ratio of the high-stress region and the rate of change of the stress gradient through an equipment degradation assessment algorithm. The pipe age parameters and material fatigue index of the water supply pipeline are obtained, and the pipeline network risk level is generated by combining them with the equipment deterioration coefficient.

Images