Evaluation system and method for state and life health of small gate hoist in irrigation area

Abstract

The invention relates to the technical field of hydraulic engineering, in particular to an irrigation area small gate hoist state and life health assessment system and method, and the system comprises a state monitoring module, a life prediction module, a fault diagnosis module, a maintenance recommendation module and a rear-end processing module which form a closed-loop collaborative architecture. The state monitoring module collects electrical parameters through a comprehensive current controller, and identifies the physical state of the gate in combination with a YOLOv8 algorithm. The life prediction module adopts a LoRA method to finely adjust the large model, and outputs a health index and residual life; the fault diagnosis module is fused with two-dimensional data to realize all-dimensional judgment; the maintenance recommendation module matches a maintenance scheme through a knowledge graph and a rule engine; and the rear end supports instruction comparison and exception notification based on a Flask framework. According to the invention, automatic and precise monitoring of the equipment state and life prediction are realized, the operation and maintenance efficiency is improved, the cost is reduced, support is provided for digital twin water conservancy, and the water supply safety of an irrigation area is guaranteed.

Description

Technical Field

[0001] This invention belongs to the field of water conservancy engineering technology, specifically relating to an assessment system and method for the status and lifespan health of small gate opening and closing mechanisms in irrigation areas. Background Technology

[0002] In irrigation water conservancy projects, gate hoists and small sluice gates are core water control equipment, and their operational status directly affects the safety of water supply and the stability of the irrigation system. Traditional irrigation areas have a large number of small sluice gates that are scattered. For a long time, the monitoring and management of the status of such equipment and gate hoists has mainly relied on manual inspection. Staff members need to conduct on-site inspections along the canals one by one: visually inspecting the equipment for obvious defects such as rust and deformation, sensing the operating temperature of the equipment by touch, measuring the tightness of the connecting parts with simple tools such as wrenches, and judging whether the equipment is faulty based on personal experience.

[0003] However, the aforementioned traditional management model has revealed many technical limitations in practical applications, making it difficult to meet the needs of modern irrigation district information and intelligent management:

[0004] (1) Lack of equipment health status assessment and lifespan prediction capabilities, resulting in passive and inefficient maintenance modes. Traditional detection methods can only preliminarily determine whether the equipment is currently in operation, and cannot predict the remaining service life of the equipment or track the trend of health status changes through data accumulation and systematic analysis. This leads to a double dilemma of "over-maintenance" and "under-maintenance" in maintenance work - either replacing equipment that can still operate normally in advance due to a lack of scientific basis, resulting in unnecessary waste of human and material resources; or failing to predict potential faults, leading to sudden equipment shutdowns, which seriously affects the stability of irrigation scheduling in the irrigation area.

[0005] (2) High reliance on manual labor makes it difficult to guarantee detection efficiency and accuracy. Due to the large number and wide distribution of small sluice gates in irrigation areas, the traditional model requires a large amount of manpower to carry out periodic inspections, which is time-consuming and labor-intensive. At the same time, the detection results are easily affected by subjective and objective conditions such as the experience level, sense of responsibility, and environmental factors of the staff. This not only makes it difficult to meet the high-frequency and full-coverage monitoring requirements, but also may lead to missed detections and misjudgments due to human negligence, resulting in low detection efficiency and inability to effectively guarantee accuracy.

[0006] (3) Lack of a digital and intelligent management system, resulting in lagging information transmission and processing. In the traditional management model, information such as equipment ledgers, operating data, and fault records are mainly transmitted and stored through paper documents and oral reports. This not only leads to significant delays in information sharing but also makes data deviations prone to human error. Furthermore, the lack of digital tools such as smart apps and remote monitoring platforms prevents staff from querying equipment-related data in real time, and management also finds it difficult to remotely monitor equipment operating status, resulting in serious deficiencies in the timeliness, convenience, and collaboration of management work.

[0007] (4) The fault detection and response mechanism is imperfect and the risk prevention and control capability is weak. Due to the limitations of the periodicity of manual inspection, equipment faults are often not detected in time. Small faults can easily deteriorate into major faults, which not only greatly increases maintenance costs, but may also cause water supply interruption in the irrigation area due to equipment shutdown, affecting agricultural production and residential water use.

[0008] In view of this, the present invention is hereby proposed. Summary of the Invention

[0009] In order to solve the above-mentioned technical problems in the prior art, the present invention provides an assessment system and method for the status and lifespan health of irrigation area small gate opening and closing mechanisms, which solves the problems of the existing irrigation area small gate and opening and closing mechanism management mode relying on manual labor, low efficiency, difficulty in predicting lifespan and delayed failure.

[0010] To achieve the above objectives, the technical solution of the present invention is as follows: In the first aspect, there is an assessment system for the status and lifespan health of a small gate in an irrigation area, comprising: a status monitoring module, a lifespan prediction module, a fault diagnosis module, a maintenance recommendation module, and a back-end processing module. The output of the status monitoring module is communicatively connected to the input of the life prediction module and the first input of the fault diagnosis module, respectively, and is used to synchronously transmit the collected electrical operation data of the hoist and the physical status data of the gate to the life prediction module and the fault diagnosis module. The output of the life prediction module is communicatively connected to the second input of the fault diagnosis module and the first input of the maintenance recommendation module, respectively, and is used to output the equipment health status assessment results and the remaining service life prediction results to the fault diagnosis module and the maintenance recommendation module. The output of the fault diagnosis module is communicatively connected to the second input of the maintenance recommendation module, and is used to output a comprehensive fault judgment result to the maintenance recommendation module. The output of the maintenance recommendation module is communicatively connected to the first input of the back-end processing module, and is used to output maintenance solutions to the back-end processing module. The backend processing module establishes bidirectional data interaction with the status monitoring module, life prediction module, fault diagnosis module, and maintenance recommendation module, respectively, to receive and store data from the above modules, coordinate the working sequence of the above modules, and output the final evaluation results.

[0011] Furthermore, the status monitoring module includes a comprehensive current controller and an image acquisition unit; The output terminals of the integrated current controller and the image acquisition unit are both communicatively connected to the first data receiving terminal of the back-end processing module. The data collected by the integrated current controller and the image acquisition unit are distributed to the life prediction module and the fault diagnosis module by the back-end processing module. The integrated current controller is used to collect the electrical operation data of the gate hoist, and the image acquisition unit is used to collect the physical state image data of the gate.

[0012] Furthermore, the integrated current controller is equipped with a high-precision sensing module, and the electrical operating data includes the effective value of voltage, harmonic distortion rate, and power factor. The integrated current controller includes local data storage. The data output terminal of the integrated current controller establishes communication with the first data receiving terminal of the back-end processing module through a wired or wireless transmission link, transmits the collected electrical operation data to the back-end processing module, and then the back-end processing module forwards it to the life prediction module.

[0013] Furthermore, the image acquisition unit is a bullet camera, and the data output end of the image acquisition unit establishes communication with the second data receiving end of the back-end processing module through a 4G / 5G or wired transmission link; The back-end processing module has a built-in visual recognition algorithm unit. The input end of the visual recognition algorithm unit is connected to the internal data interface of the back-end processing module. It is used to receive physical state image data and perform frame extraction processing and feature analysis to identify the opening and closing status of the gate, surface defects and structural integrity information. The recognition results of the visual recognition algorithm unit are transmitted to the first input end of the fault diagnosis module through the back-end processing module.

[0014] Furthermore, the visual recognition algorithm unit employs the YOLOv8 algorithm, and the recognition process of the YOLOv8 algorithm includes: Detect the boundary frames of the gate body and the gate frame, and calculate their intersection-union ratio. The formula for calculating the intersection-union ratio is:

[0015] in, The bounding box of the target region. For reference bounding box; Detect the rusted areas on the gate surface and calculate the percentage of the rusted area's pixel area to the total pixel area of ​​the gate. By combining key point detection with gate corner point positioning, the deviation of the angle between opposite side vectors from 90 degrees is calculated. Alternatively, use the Hausdorff distance to analyze the geometric relationships between the gate corners or the differences between the actual gate profile and the standard rectangle, and the deviation. The calculation formula is:

[0016] The output of the YOLOv8 algorithm is transmitted to the fault diagnosis module as the basis for determining the physical faults of the gate.

[0017] Furthermore, the life prediction module has a built-in large model unit. The input of the large model unit receives electrical operation data transmitted by the status monitoring module, and the output of the large model unit outputs the equipment health index and remaining service life. The large model unit is fine-tuned based on a structured dataset with paired features and labels. The features of the structured dataset include: total harmonic distortion of current, total harmonic distortion of voltage, cumulative start-stop count, cumulative runtime, effective voltage value, and power factor. The label of the structured dataset is the device health index.

[0018] Furthermore, the fine-tuning of the large model unit employs the LoRA parameter efficient fine-tuning method, injecting a LoRA adapter through the PEFT library. The forward propagation formula for the LoRA parameter efficient fine-tuning method is as follows:

[0019] in, For forward propagation output, For pre-trained weights, It is a low-rank matrix; When preprocessing electrical operation data, the Fast Fourier Transform algorithm is used to calculate the total harmonic distortion (THD). The formula for calculating the THD is:

[0020] This is the sum of the squares of the effective values ​​of the voltages of all harmonics from the second harmonic to the highest harmonic considered. This is the effective value of the fundamental voltage; The large model unit is trained using the AdamW optimizer. During training, the rank and lora_alpha hyperparameters are adjusted, and the appropriate query / value layers are specified. The health index and remaining service life output by the fine-tuned large model unit are transmitted to the fault diagnosis module and the maintenance recommendation module.

[0021] Furthermore, the fault diagnosis module has a built-in fusion logic unit. The first input of the fusion logic unit receives the recognition result from the visual recognition algorithm unit, and the second input of the fusion logic unit receives the health status assessment result from the large model unit. Through cross-validation using rule and probability fusion logic, the root cause of the fault, the fault location, and the fault confidence are determined. The fault judgment result of the fault diagnosis module is transmitted to the maintenance recommendation module.

[0022] Furthermore, the maintenance recommendation module incorporates a knowledge graph unit and a rule engine unit. The knowledge graph unit stores historical fault maintenance data, and the rule engine unit receives the fault judgment results from the fault diagnosis module and the health index and remaining service life from the life prediction module. By matching historical fault maintenance data with real-time data, it outputs a maintenance plan that includes maintenance operation steps, maintenance priority, and maintenance timing. The maintenance plan from the maintenance recommendation module is then transmitted to the backend processing module.

[0023] Furthermore, the backend processing module is developed based on the Flask framework of Python. The backend processing module obtains the gate opening and closing instruction information through the official port and compares it with the gate opening and closing status identified by the visual recognition algorithm unit to determine whether there is an instruction failure. When the fault diagnosis module determines that the equipment is faulty or the equipment health index is lower than the preset value, the backend processing module sends a notification message to the preset contact person via telephone or SMS API, and outputs the equipment health assessment report and maintenance plan.

[0024] Secondly, a method for assessing the status and lifespan health of small gate opening and closing mechanisms in irrigation districts includes: S1. The electrical operation data of the hoist and the physical status data of the gate are acquired simultaneously through the acquisition equipment to form dual-dimensional monitoring data; S2. Based on the preset structured dataset, perform domain fine-tuning on the general large model, and use the fine-tuned model to receive the electrical operation data obtained in the first step, and output the equipment health status assessment results and the remaining service life prediction results. S3. Integrate the dual-dimensional monitoring data with the health status assessment results and remaining service life prediction results to make a comprehensive judgment on equipment failure; S4. Based on the fault determination results, health status assessment results, and remaining service life prediction results, and combined with historical equipment data, a maintenance plan is matched and generated. S5. Receive the repair plan and coordinate the work sequence, and output the final evaluation result.

[0025] Compared with existing technologies, the present invention provides an assessment system and method for the status and lifespan health of small gate hoists in irrigation areas. The system includes a status monitoring module, a lifespan prediction module, a fault diagnosis module, a maintenance recommendation module, and a back-end processing module. These modules form a closed-loop collaborative architecture through a data transmission link. The status monitoring module collects electrical parameters through a comprehensive current controller and identifies the physical status of the gate using the YOLOv8 image algorithm. The lifespan prediction module, based on a feature- and label-structured dataset, fine-tunes a general large model using the LoRA method, outputting a health index and remaining lifespan. The fault diagnosis module integrates visual recognition results with electrical assessment results to achieve comprehensive fault determination. The maintenance recommendation module matches the optimal maintenance plan using a knowledge graph and rule engine. The back-end processing module, developed based on the Flask framework, supports instruction comparison, anomaly notification, and result output. This invention achieves automated and precise monitoring and lifespan prediction of equipment status, improves operation and maintenance efficiency, reduces costs, provides support for digital twin water conservancy, and ensures the safety of water supply in irrigation areas. Attached Figure Description

[0026] Figure 1 This is an architecture diagram of the irrigation district small gate opening and closing mechanism status and lifespan health assessment system provided in an embodiment of the present invention. Figure 2 A flowchart illustrating the method for assessing the status and lifespan health of small gate opening and closing mechanisms in irrigation areas, as provided in an embodiment of the present invention. Detailed Implementation

[0027] The technical solution of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0028] It should be noted that, unless otherwise specifically stated, the relative arrangement and numerical expressions of the components and steps described in these embodiments should not be construed as limiting the scope of the invention.

[0029] The following description of exemplary embodiments is merely illustrative and is not intended to limit the invention or its application or use in any way. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail herein, but where applicable, such techniques, methods, and apparatus should be considered part of this specification.

[0030] Example 1 See Figure 1 , Figure 1 The present invention provides an architecture diagram of an assessment system for the status and lifespan health of a small gate opening and closing mechanism in an irrigation district, comprising: M1, Status Monitoring Module: Its output terminal is communicatively connected to the input terminal of the life prediction module and the first input terminal of the fault diagnosis module, respectively, for synchronously transmitting the collected electrical operation data of the hoist and the physical status data of the gate to the life prediction module and the fault diagnosis module; specifically including: M11, Integrated Current Controller: Used to collect electrical operation data of the hoist; all output terminals are connected to the first data receiving terminal of the back-end processing module. The data from the integrated current controller is distributed to the life prediction module and the fault diagnosis module by the back-end processing module. The integrated current controller uses a high-precision sensing module to capture key electrical parameters such as the effective voltage value, harmonic distortion rate, and power factor of the hoist in real time. It has a fast response capability and can accurately reflect the power fluctuations during the operation of the hoist. It also supports local data storage and real-time transmission. The data output end of the integrated current controller establishes communication with the first data receiving end of the back-end processing module through a wired or wireless transmission link, transmits the collected electrical operation data to the back-end processing module, and then the back-end processing module forwards it to the life prediction module.

[0031] M12, Image Acquisition Unit: Used to acquire physical state image data of the gate; the output end of the image acquisition unit is connected to the first data receiving end of the back-end processing module. The data acquired by the image acquisition unit is distributed to the life prediction module and the fault diagnosis module by the back-end processing module. The image acquisition unit is a bullet camera. The data output end of the image acquisition unit establishes communication with the second data receiving end of the back-end processing module through a 4G / 5G or wired transmission link. The back-end processing module has a built-in visual recognition algorithm unit. The input end of the visual recognition algorithm unit is connected to the internal data interface of the back-end processing module. It is used to receive physical state image data and perform frame extraction processing and feature analysis to identify the opening and closing status of the gate, surface defects and structural integrity information. The recognition result of the visual recognition algorithm unit is transmitted to the first input end of the fault diagnosis module through the back-end processing module.

[0032] The visual recognition algorithm unit adopts the YOLOv8 algorithm, which extracts frames from the video stream and feeds them into the YOLOv8 model for inference to realize the recognition of the physical state of the gate.

[0033] The YOLOv8 model is pre-trained using a dataset labeled with targets such as gates, rust, and deformation to learn relevant visual features. During recognition, it detects the bounding boxes of the gate body and the gate frame, calculates the intersection-over-union (IoU) ratio between them to determine the opening and closing status. The IoU ratio is calculated using the following formula:

[0034] in, For example, the bounding box of the target area automatically identified by the algorithm represents the coordinate range of a specific target object detected from video surveillance in irrigation district applications. The true values ​​of the data are manually and precisely labeled to serve as reference bounding boxes; when The gate is considered open when the threshold is reached; the degree of corrosion is assessed by statistically analyzing the proportion of the area of ​​corroded pixels to the total gate area. When the corrosion proportion is greater than a threshold... The corresponding alarm level is triggered in a timely manner; the gate corner point is located by combining key point detection, and the deviation of the included angle of the opposite side vector from 90 degrees is calculated. Alternatively, use the Hausdorff distance to analyze the geometric relationships between the gate corners or the differences between the actual gate profile and the standard rectangle, and the deviation. The calculation formula is:

[0035] when Deformation is determined in real time, and the identification result is transmitted to the first input terminal of the fault diagnosis module through the back-end processing module; the output result of the YOLOv8 algorithm is used as the basis for determining the physical fault of the gate and transmitted to the fault diagnosis module.

[0036] M2, Lifespan Prediction Module: Its output is communicatively connected to the second input of the fault diagnosis module and the first input of the maintenance recommendation module, respectively, and is used to output equipment health status assessment results and remaining service life prediction results to the fault diagnosis module and the maintenance recommendation module; specifically including: M21, Large Model Unit: The lifetime prediction module has a built-in large model unit that has been fine-tuned in the domain. This large model unit is fine-tuned based on a structured dataset with paired features and labels. The features of the structured dataset include total harmonic distortion of current, total harmonic distortion of voltage, cumulative start-stop count, cumulative runtime, effective voltage value and power factor, and the label is the Equipment Health Index (HI).

[0037] The fine-tuning of large model units employs the LoRA parameter efficiency fine-tuning method, which injects a LoRA adapter through the PEFT library. The forward propagation formula for the LoRA parameter efficiency fine-tuning method is as follows:

[0038] in, For forward propagation output, For pre-trained weights, It is a low-rank matrix; After preprocessing the raw voltage and current time-series data collected by the integrated current controller, the total harmonic distortion (THD) is calculated using the Fast Fourier Transform (FFT) algorithm during the preprocessing of electrical operation data. The formula for calculating the THD is as follows:

[0039] in, This is the sum of the squares of the effective values ​​of the voltages of all harmonics from the second harmonic to the highest harmonic considered. This is the effective value of the fundamental voltage; The large model unit is trained using the AdamW optimizer. During training, the rank and lora_alpha hyperparameters are adjusted, and the appropriate query / value layers are specified. The health index and remaining service life output by the fine-tuned large model unit are transmitted to the fault diagnosis module and the maintenance recommendation module.

[0040] The preprocessed electrical operation data is then input into the fine-tuned large model unit, which outputs the Equipment Health Index (HI) and Remaining Service Life (RUL). This output is simultaneously transmitted to the second input of the fault diagnosis module and the first input of the maintenance recommendation module.

[0041] M3, Fault Diagnosis Module: Its output terminal communicates with the second input terminal of the maintenance recommendation module, and is used to output comprehensive fault diagnosis results to the maintenance recommendation module; specifically including: M31, Fusion Logic Unit: The first input terminal receives the recognition result of the visual recognition algorithm unit, and the second input terminal of the fusion logic unit receives the health status assessment result of the large model unit. Through rule and probability fusion logic, cross-validation is performed to determine the root cause of the fault, the fault location, and the fault confidence. The fault judgment result of the fault diagnosis module is transmitted to the maintenance recommendation module.

[0042] When the visual recognition algorithm identifies physical defects such as rust or deformation of the gate or abnormal opening, and the large model unit outputs the internal fault probability and health score after analyzing the electrical parameters, the fusion logic unit cross-validates the two types of results. If the visual recognition detects mechanical jamming and the electrical model detects overcurrent characteristics, it is diagnosed as "mechanical jamming causing overload". Finally, a fault judgment result containing the fault root cause, location, confidence level and maintenance priority is generated and transmitted to the second input terminal of the maintenance recommendation module.

[0043] M4, Repair Recommendation Module: The output end is connected to the first input end of the back-end processing module, and is used to output repair solutions to the back-end processing module; specifically including: M41, Knowledge Graph Unit: Used to store historical fault repair data; M42, Rule Engine Unit: The input end receives the fault judgment results from the fault diagnosis module, the health index (HI) and remaining service life (RUL) from the life prediction module, and the rule engine. Based on multi-dimensional real-time monitoring data (electrical parameters, physical status, operating condition information) and historical fault data, the optimal maintenance plan is matched through knowledge graph and rule engine, and the maintenance timing is recommended in combination with the life prediction results. The maintenance plan containing maintenance operation steps is generated and transmitted to the back-end processing module.

[0044] M5, Backend Processing Module: Establishes bidirectional data interaction with the Status Monitoring Module, Life Prediction Module, Fault Diagnosis Module, and Maintenance Recommendation Module respectively. It is used to receive and store data from the above modules, coordinate the working sequence of the above modules, and output the final evaluation results.

[0045] The backend processing module is developed based on the Flask framework in Python. It obtains gate opening and closing command information through the official port and compares it with the gate opening and closing status identified by the visual recognition algorithm unit to determine whether there is any command failure. The backend processing module establishes bidirectional data interaction with the status monitoring module, life prediction module, fault diagnosis module, and maintenance recommendation module, respectively, receives and stores data from each module, and coordinates the working sequence of each module. When the fault diagnosis module determines that the equipment has a fault or the equipment health index (HI) is lower than the preset value, the backend processing module sends a notification message to the preset contact via telephone or SMS API, and outputs an equipment health assessment report and maintenance plan.

[0046] The system's workflow includes: the integrated current controller and the camera in the status monitoring module simultaneously collect electrical operation data of the gate hoist and video streams of the gate's physical status, and upload them to the back-end processing module via 4G / 5G or wired transmission links; the back-end processing module forwards the electrical operation data to the life prediction module, and transmits the video stream to the YOLOv8 algorithm unit for processing, obtaining the health index, remaining life prediction results, and gate physical status identification results respectively; the fault diagnosis module integrates the above two types of results, and outputs the fault judgment result after cross-validation by the fusion logic unit; the maintenance recommendation module matches and generates maintenance plans based on the fault judgment results, health index, remaining life, and historical fault data, and transmits them to the back-end processing module; the back-end processing module coordinates the working sequence of each module, stores all relevant data, outputs equipment health assessment reports and maintenance plans, and sends notification information in case of abnormalities, realizing a full-process intelligent assessment of the status and life health of the small gate hoist in the irrigation area.

[0047] Example 2 See Figure 2 , Figure 2 The flowchart of the method for assessing the status and lifespan health of small gate opening and closing mechanisms in irrigation areas, as proposed in this invention, includes: S1. The electrical operation data of the hoist and the physical status data of the gate are acquired simultaneously through the acquisition equipment to form dual-dimensional monitoring data; S2. Based on the preset structured dataset, perform domain fine-tuning on the general large model, and use the fine-tuned model to receive the electrical operation data obtained in the first step, and output the equipment health status assessment results and the remaining service life prediction results. S3. Integrate the dual-dimensional monitoring data with the health status assessment results and remaining service life prediction results to make a comprehensive judgment on equipment failure; S4. Based on the fault determination results, health status assessment results, and remaining service life prediction results, and combined with historical equipment data, a maintenance plan is matched and generated. S5. Receive the repair plan and coordinate the work sequence, and output the final evaluation result.

[0048] In summary, the present invention has the following advantages: 1. Outstanding risk prediction capabilities and enhanced safety assurance: Relying on the LoRA fine-tuned water conservancy expert big model, it can identify early signs of equipment abnormalities, and judge the health status of equipment and predict its life by analyzing voltage and current data. This helps managers to take intervention measures before substantial equipment damage, reduce the occurrence of safety accidents, reduce economic losses caused by failures, and ensure the safe operation of irrigation area water conservancy facilities. 2. The status monitoring combines comprehensiveness and accuracy: It adopts a monitoring method that combines large language models and image vision algorithms. On the one hand, it detects abnormalities in the appearance of the gate through image algorithms, and on the other hand, it monitors the electrical data of the hoist, so as to realize multi-dimensional coverage monitoring of equipment status and improve the accuracy of monitoring results. 3. Improved automation and optimized operation and maintenance efficiency: Replacing traditional manual periodic inspections with automated testing significantly reduces labor costs; at the same time, it optimizes the allocation of maintenance resources, reduces overall operation and maintenance costs, and significantly improves work efficiency and resource utilization.

[0049] The above specific embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A system for assessing the status and lifespan health of small gate opening and closing mechanisms in irrigation districts, characterized in that, include: Condition monitoring module, life prediction module, fault diagnosis module, maintenance recommendation module and back-end processing module; The output of the status monitoring module is communicatively connected to the input of the life prediction module and the first input of the fault diagnosis module, respectively, and is used to synchronously transmit the collected electrical operation data of the hoist and the physical status data of the gate to the life prediction module and the fault diagnosis module. The output of the life prediction module is communicatively connected to the second input of the fault diagnosis module and the first input of the maintenance recommendation module, respectively, and is used to output the equipment health status assessment results and the remaining service life prediction results to the fault diagnosis module and the maintenance recommendation module. The output of the fault diagnosis module is communicatively connected to the second input of the maintenance recommendation module, and is used to output a comprehensive fault judgment result to the maintenance recommendation module. The output of the maintenance recommendation module is communicatively connected to the first input of the back-end processing module, and is used to output maintenance solutions to the back-end processing module. The backend processing module establishes bidirectional data interaction with the status monitoring module, life prediction module, fault diagnosis module, and maintenance recommendation module, respectively, to receive and store data from the above modules, coordinate the working sequence of the above modules, and output the final evaluation results.

2. The assessment system for the status and lifespan health of small gate opening and closing mechanisms in irrigation areas according to claim 1, characterized in that, The status monitoring module includes a comprehensive current controller and an image acquisition unit; The output terminals of the integrated current controller and the image acquisition unit are both communicatively connected to the first data receiving terminal of the back-end processing module. The data collected by the integrated current controller and the image acquisition unit are distributed to the life prediction module and the fault diagnosis module by the back-end processing module. The integrated current controller is used to collect the electrical operation data of the gate hoist, and the image acquisition unit is used to collect the physical state image data of the gate.

3. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 2, characterized in that, The integrated current controller is equipped with a high-precision sensing module, and the electrical operating data includes the effective value of voltage, harmonic distortion rate and power factor; The integrated current controller includes local data storage. The data output terminal of the integrated current controller establishes communication with the first data receiving terminal of the back-end processing module through a wired or wireless transmission link, transmits the collected electrical operation data to the back-end processing module, and then the back-end processing module forwards it to the life prediction module.

4. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 2, characterized in that, The image acquisition unit is a bullet camera, and the data output end of the image acquisition unit establishes communication with the second data receiving end of the back-end processing module through a 4G / 5G or wired transmission link. The back-end processing module has a built-in visual recognition algorithm unit. The input end of the visual recognition algorithm unit is connected to the internal data interface of the back-end processing module. It is used to receive physical state image data and perform frame extraction processing and feature analysis to identify the opening and closing status of the gate, surface defects and structural integrity information. The recognition results of the visual recognition algorithm unit are transmitted to the first input end of the fault diagnosis module through the back-end processing module.

5. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 4, characterized in that, The visual recognition algorithm unit uses the YOLOv8 algorithm, and the recognition process of the YOLOv8 algorithm includes: Detect the boundary frames of the gate body and the gate frame, and calculate their intersection-union ratio. The formula for calculating the intersection-union ratio is: in, The bounding box of the target region. For reference bounding box; Detect the rusted areas on the gate surface and calculate the percentage of the rusted area's pixel area to the total pixel area of ​​the gate. By combining key point detection with gate corner point positioning, the deviation of the angle between opposite side vectors from 90 degrees is calculated. Alternatively, use the Hausdorff distance to analyze the geometric relationships between the gate corners or the differences between the actual gate profile and the standard rectangle, and the deviation. The calculation formula is: The output of the YOLOv8 algorithm is transmitted to the fault diagnosis module as the basis for determining the physical faults of the gate.

6. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 1, characterized in that, The life prediction module has a built-in large model unit. The input of the large model unit receives electrical operation data transmitted by the status monitoring module, and the output of the large model unit outputs the equipment health index and remaining service life. The large model unit is fine-tuned based on a structured dataset with paired features and labels. The features of the structured dataset include: total harmonic distortion of current, total harmonic distortion of voltage, cumulative start-stop count, cumulative runtime, effective voltage value, and power factor. The label of the structured dataset is the device health index.

7. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 6, characterized in that, The fine-tuning of the large model units employs the LoRA parameter-efficient fine-tuning method, which injects a LoRA adapter through the PEFT library. The forward propagation formula for the LoRA parameter-efficient fine-tuning method is as follows: in, For forward propagation output, For pre-trained weights, It is a low-rank matrix; When preprocessing electrical operation data, the Fast Fourier Transform algorithm is used to calculate the total harmonic distortion (THD). The formula for calculating the THD is: This is the sum of the squares of the effective values ​​of the voltages of all harmonics from the second harmonic to the highest harmonic considered. This is the effective value of the fundamental voltage; The large model unit is trained using the AdamW optimizer. During training, the rank and lora_alpha hyperparameters are adjusted, and the appropriate query / value layers are specified. The health index and remaining service life output by the fine-tuned large model unit are transmitted to the fault diagnosis module and the maintenance recommendation module.

8. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 1, characterized in that, The fault diagnosis module has a built-in fusion logic unit. The first input of the fusion logic unit receives the recognition result of the visual recognition algorithm unit, and the second input of the fusion logic unit receives the health status assessment result of the large model unit. Through cross-validation using rule and probability fusion logic, the root cause of the fault, the fault location, and the fault confidence are determined. The fault judgment result of the fault diagnosis module is transmitted to the maintenance recommendation module.

9. The assessment system for the status and lifespan health of small gate opening and closing mechanisms in irrigation areas according to claim 1, characterized in that, The maintenance recommendation module has a built-in knowledge graph unit and a rule engine unit. The knowledge graph unit stores historical fault maintenance data. The input of the rule engine unit receives the fault judgment results from the fault diagnosis module and the health index and remaining service life from the life prediction module. By matching historical fault maintenance data with real-time data, it outputs a maintenance plan that includes maintenance operation steps, maintenance priority, and maintenance timing. The maintenance plan of the maintenance recommendation module is transmitted to the back-end processing module.

10. The assessment system for the status and lifespan health of irrigation district small gate opening and closing mechanisms according to claim 1, characterized in that, The backend processing module is developed based on the Flask framework of Python. The backend processing module obtains the gate opening and closing instruction information through the official port and compares it with the gate opening and closing status identified by the visual recognition algorithm unit to determine whether there is an instruction failure. When the fault diagnosis module determines that the equipment is faulty or the equipment health index is lower than the preset value, the backend processing module sends a notification message to the preset contact person via telephone or SMS API, and outputs the equipment health assessment report and maintenance plan.

11. A method for assessing the status and lifespan health of small gate opening and closing mechanisms in irrigation districts, characterized in that... include: S1. The electrical operation data of the hoist and the physical status data of the gate are acquired simultaneously through the acquisition equipment to form dual-dimensional monitoring data; S2. Based on the preset structured dataset, perform domain fine-tuning on the general large model, and use the fine-tuned model to receive the electrical operation data obtained in the first step, and output the equipment health status assessment results and the remaining service life prediction results. S3. Integrate the dual-dimensional monitoring data with the health status assessment results and remaining service life prediction results to make a comprehensive judgment on equipment failure; S4. Based on the fault determination results, health status assessment results, and remaining service life prediction results, and combined with historical equipment data, a maintenance plan is matched and generated. S5. Receive the repair plan and coordinate the work sequence, and output the final evaluation result.

Images