Intelligent water washing method based on insulator smudginess degree

Abstract

The invention discloses an intelligent water washing method based on insulator smudginess degree, and the method specifically comprises the steps: carrying on-line insulator smudginess detection sensor equipment through a water washing vehicle, automatically collecting and analyzing the smudginess degree of an insulator on a railway overhead line system line in real time, and judging the leakage current risk of the insulator; the insulators are automatically identified and positioned through multi-sensor equipment with vision and laser radar fused, and the shapes and sizes of the insulators are distinguished; then differentiated flushing strategies are automatically produced, targeted deep cleaning is carried out through an intelligent water flushing system, and after flushing, detection and judgment are carried out again through online insulator dirt detection sensor equipment carried at the other end, so that an operation closed loop is formed. According to the invention, the cleaning mode from regular general cleaning to on-demand fine cleaning is converted, the defects that the existing operation mode depends on artificial experience, blind flushing is caused and water resources are wasted are overcome, and fine, intelligent and digital management of insulator cleaning is realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of maintenance of insulators of electrified railway catenary lines, and particularly relates to an intelligent water flushing method based on the dirtiness degree of insulators. BACKGROUND

[0002] In electrified railways, catenary insulators play an important role in support and insulation. However, due to long-term exposure to outdoor environments, dust, dirt and other pollutants can accumulate on the surface of the insulators, which can reduce the insulation performance of the insulators and even cause pollution flashover accidents, endangering the safety of train operation. Therefore, regular cleaning of insulators is one of the important measures to ensure the safe operation of electrified railways. Traditional cleaning methods such as manual cleaning, mechanical cleaning or water flushing have problems such as time-consuming, labor-intensive, high cost or geographical limitations. Currently, the main cleaning method for insulators is live water flushing. Both live water flushing and other water flushing methods mainly rely on human experience and fixed flushing plans. The operation mainly has the following problems:

[0003] Blindness: The actual dirtiness of the insulators cannot be sensed in real time, and a "one-size-fits-all" flushing scheme is often used, which either results in safety hazards due to insufficient flushing or wastes water resources and manpower due to excessive flushing.

[0004] Lag: The planned flushing based on fixed periods cannot respond to rapid contamination of insulators caused by sudden pollution events such as mine pollution, factory leaks and sandstorms.

[0005] Lack of quantitative basis: The flushing effect is difficult to quantify and evaluate in real time, relying on the naked eye observation of the operator, which is highly subjective and not precise in management.

[0006] Even though there are intelligent flushing equipment and online insulator dirtiness detection devices in development and trial operation, there is still no universal and formal deployment and formation of routine operation procedures, and it still belongs to independent and fragmented technical modules. However, in 2021, the team from Southwest Jiaotong University first proposed "an intelligent dirtiness degree judgment insulator flushing system (patent number: CN202111598553.8)", which mainly describes a technology for detecting the dirtiness of insulators through single camera imaging comparison and automatic water cannon linkage flushing, without forming a complete "detection-cleaning-evaluation" closed-loop operation process method and post-flushing effect evaluation, with unknown results and no data retention. Therefore, there is an urgent need to integrate technologies and complete devices to form a "one-stop" insulator flushing method that can sense in real time, make intelligent decisions, execute accurately and evaluate effects, and establish a digital line network to realize the fine, digital and intelligent operation of catenary line maintenance. SUMMARY

[0007] The present application aims to be able to evaluate the contamination state and risk of insulators in real time, accurately, and automatically perform fine flushing operations matching the same, timely judge the operation effect, long-term establish line insulator contamination condition big data, so as to realize safe, efficient and energy-saving line maintenance. For this purpose, the present application provides an intelligent water flushing method based on the contamination degree of insulators.

[0008] The core of the intelligent water flushing method based on the contamination degree of insulators of the present application is to build a closed-loop intelligent control process of "perception-evaluation-decision-execution-verification". Specifically, the following steps are included:

[0009] Step 1: Insulator identification, positioning and tracking.

[0010] Visual camera, laser radar and Beidou positioning are carried on the front end and the rear end of the water flushing vehicle, and the identification and positioning of insulators and pole numbers are carried out through multi-sensor fusion technology, so as to control and guide the automatic water cannon to track and aim at the insulator target in real time, and be ready to flush the insulator at any time. The front and rear are carried for the convenience of bidirectional operation of the water flushing vehicle.

[0011] Step 2: Generating insulator database.

[0012] After identifying the insulators and the pole numbers, the line map coordinates are drawn through the Beidou positioning system, and then the corresponding pole numbers and insulator numbers identified and positioned are input into the line map, so as to form a line network database.

[0013] Step 3: Real-time collection of insulator surface contamination data.

[0014] Through the insulator surface contamination detection equipment (using optical sensor, high-definition camera and ultrasonic leakage current monitoring device) carried on the water flushing vehicle, the surface state parameters and leakage current parameters of the target insulator are obtained in real time in a non-contact manner.

[0015] Step 4: Dynamic evaluation of insulator leakage current risk.

[0016] The surface state parameters and leakage current parameters are processed, the comprehensive contamination index of the insulator is calculated through the contamination evaluation model (a machine learning model trained by historical data in the early stage), and the current risk level of the insulator is dynamically evaluated by fusing the leakage current parameters.

[0017] Step 5: Generating insulator flushing strategy.

[0018] According to the current risk level and the shape and size of the insulator, the corresponding flushing control instructions are matched and generated from the pre-defined flushing strategy library, and the flushing control instructions at least define the flushing water pressure, the flushing duration and the trajectory of the flushing jet scanning the insulator.

[0019] Step 6: Perform a precise rinse.

[0020] The actuators of the water flushing vehicle (such as variable frequency water pumps, solenoid valves, and servo-driven water cannons) receive and analyze the above flushing control commands, automatically adjust to the state required by the commands, and aim at the target insulator to complete the entire flushing operation. The entire process requires no manual intervention in the flushing parameters.

[0021] Furthermore, it also includes the following steps:

[0022] Step 7: Verify the rinsing effect.

[0023] After the washing operation is completed, steps 3 and 4 are executed again by the insulator surface dirt detection equipment at the other end of the vehicle to obtain the status parameters of the insulator after washing and assess its risk level. If the comprehensive dirt index after washing is lower than the safety threshold and the risk level drops to low, the washing is deemed qualified; otherwise, the washing is deemed unqualified, and the insulator data is listed as a warning. Based on the current detection results, the washing strategy is optimized (such as increasing the pressure or extending the time) and a second washing is performed, or a second washing is performed on the return trip of the vehicle, until the verification is qualified.

[0024] Furthermore, in step 3, the surface state parameters are obtained through optical sensing or image recognition technology, including estimated values ​​of equivalent salt density (ESDD) and equivalent gray density (NSDD).

[0025] Leakage current parameters are obtained using ultrasonic technology, including the effective value of the leakage current and the pulse frequency.

[0026] Furthermore, in step 4, the dirt assessment model is a model trained based on machine learning algorithms. Its input features include ESDD, NSDD, ambient humidity, and temperature, and its output is a comprehensive dirt index.

[0027] The specific process for dynamically assessing risk levels is as follows: A weighted fusion analysis is performed on the comprehensive contamination index, the effective value of the leakage current, and the pulse frequency. Based on a preset threshold range, the risk level is divided into five levels: high, medium, low, and relatively low. For example, a high or high contamination index accompanied by a high leakage current pulse is considered high risk.

[0028] Furthermore, in step 5, the predefined flushing strategy library contains flushing strategies for different risk levels as follows:

[0029] For higher and higher risk levels, instructions are generated to use a powerful flushing mode with high water pressure, large water volume, and long duration.

[0030] For medium-risk levels, instructions are generated to use a standard flushing mode with medium water pressure and volume.

[0031] For low-risk levels, an instruction is generated to use an energy-saving flushing mode with low water pressure and small water volume.

[0032] For lower risk levels, generate an instruction to temporarily not flush and only record the status.

[0033] Meanwhile, different sweeping trajectories are used to clean insulators of different shapes and sizes: horizontal bracket insulators are cleaned by left and right sweeping and up and down vibration; oblique bracket insulators are cleaned by stepped up and down sweeping; and return wire insulators are cleaned by left and right swinging.

[0034] Furthermore, it also includes historical data learning: collecting and learning historical flushing data, environmental data, and insulator fault data to dynamically optimize the parameters of the dirt assessment model and the flushing strategies in the predefined flushing strategy library.

[0035] Furthermore, it also includes uploading all data generated in steps 1-7, including detection data, risk levels, flushing parameters and verification results, to the remote management platform to form a digital archive of the insulator's condition and maintenance throughout its entire lifecycle, for subsequent analysis and decision support.

[0036] The beneficial technical effects of this invention compared to the prior art are as follows:

[0037] 1. Precision and efficiency: It realizes "washing according to dirt", and performs differentiated washing according to the actual degree of dirt and risk level of the insulator, avoiding waste of resources and blind spots in washing, and significantly improving the efficiency and quality of washing operations.

[0038] 2. Proactive Safety: By monitoring leakage current in real time, potential flashover risks can be detected in a timely manner, and powerful cleaning can be prioritized. The "post-flushing verification" mechanism ensures the quality of flushing, forming a complete quality control closed loop, eliminating the uncertainty of flushing effect, and preventing accidents from happening in the first place, greatly improving the safety of train operation.

[0039] 3. Water-saving and environmentally friendly: It avoids unnecessary rinsing, especially when using energy-saving mode for low-risk insulators or only for monitoring, which can save a lot of precious water resources.

[0040] 4. Digital Management: Data throughout the entire process is recorded and can be uploaded to the cloud platform, forming a "digital profile" of the insulator's condition. This provides a verifiable big data foundation for condition-based maintenance, life prediction, and full life cycle management of power lines, and promotes the digital transformation of power transmission operation and maintenance. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the overall process of the intelligent water flushing method based on the degree of dirtiness of insulators according to the present invention.

[0042] Figure 2 This is an example diagram of an intelligent water washing vehicle based on the degree of dirtiness of insulators according to the method of the present invention.

[0043] In the diagram: 1-Front-end insulator and pole number identification device, 2-Beidou positioning device, 3-Front-end insulator pollution detection device, 4-Control and electrical control room, 5-Automatic water cannon No. I, 6-Water pressure pump set, 7-Water storage tank, 8-Automatic water cannon No. II, 9-Rear-end insulator pollution detection device, 10-Front-end insulator and pole number identification device, 11-Generator room, 12-Contact network insulator, 12-Pole number plate. Detailed Implementation

[0044] The method of the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0045] Traditional water flushing methods suffer from limitations such as inaccuracies, delays, and a lack of quantitative data. With the increasing maturity of intelligent water flushing equipment and the continuous upgrading of various online insulator contamination detection devices supported by more data, intelligent railway catenary maintenance has become imperative. Therefore, this invention integrates various emerging technologies to propose a novel intelligent water flushing method based on the degree of insulator contamination, bringing new hope to the development of a digital and intelligent maintenance system for railway catenaries.

[0046] This invention integrates an online insulator dirt detection sensor module, a data processing and risk assessment module, an intelligent water flushing control module, an effect verification and feedback module, and a database establishment, storage, and uploading module into an existing electrified water flushing vehicle. Through data sharing and interconnection, a digital network is established, forming a complete closed-loop operation process and method of "detection-cleaning-evaluation." This contributes to the construction of a refined, digital, and intelligent railway operation and maintenance system.

[0047] Reference Figure 2 An example diagram of an intelligent water washing vehicle based on the degree of dirtiness of insulators is shown. This embodiment demonstrates a complete intelligent water washing operation process, such as... Figure 1 As shown, the specific steps include:

[0048] Step 1, Start-up and Identification: After the water washing truck arrives at the starting point of the operation, the generator is started, the control system is powered on and started, all equipment systems are activated and begin to work, the vehicle moves forward at a speed of 10km / h, and the intelligent insulator identification and positioning system at the front of the vehicle begins to detect and identify the target insulator and locate and output its dynamic coordinate value.

[0049] Step 2, Database Generation or Update: By using BeiDou to locate the vehicle's position on the line, an insulator database is established or queried based on the identified insulator pole numbers. The current insulator database is then activated, awaiting the writing or updating of detection data.

[0050] Step 3, Data Acquisition Before Washing: As the vehicle moves forward, the non-contact insulator dirt detection device, immediately following the insulator identification and positioning system, moves beneath the insulator and completes the scanning and data acquisition of the target insulator within seconds. Assume the acquired data is: ESDD = 0.08 mg / cm², NSDD = 0.4 mg / cm², effective leakage current of 12mA, and pulse frequency of 8 times / minute. This data is then stored, written to, or updated in the corresponding insulator database.

[0051] Step 4, Risk Assessment: The data processing unit then inputs the above data along with the ambient temperature and humidity data into the pre-trained Gradient Boosting Decision Tree (GBDT) model, calculating the comprehensive contamination index FI = 0.55. Subsequently, the risk assessment logic is activated: Since the FI value is in the medium-risk range (0.3, 0.6] and the leakage current pulse frequency is low, the system ultimately determines the insulator to be a "horizontal brace insulator, Level III medium risk".

[0052] Step 5, Strategy Generation: Based on the "medium risk" level, the system calls the corresponding standard flushing mode from the strategy library and generates the following flushing instruction: {Adjust the water pump motor frequency to 40Hz; water pressure: 1.5MPa; continuous water jet time: 6 seconds; perform horizontal sweeping and vertical vibration}.

[0053] Step 6, Flushing Execution: The water pressure control system and water cannon servo system of the water flushing truck receive the command, automatically adjust the water pump pressure to 1.5MPa, control the water spray volume to 15L / min, and automatically aim and start continuous tracking and sweeping flushing. It will automatically stop after 6 seconds.

[0054] Step 7, Effect Verification: After flushing stops, the vehicle continues to move forward. The non-contact insulator dirt detection device at the other end of the vehicle moves under the insulator, and the system scans and detects the insulator again. Assume the collected data is: ESDD = 0.02 mg / cm², NSDD = 0.05 mg / cm², leakage current drops to 2mA with no pulse. The calculated FI value drops to 0.08, far below the safety threshold (e.g., 0.15). The system determines "flushing qualified," and the operation for this insulator identification pole number is completed, moving on to the next insulator identification pole number. All data (including before-and-after comparison data) is packaged and uploaded to the cloud platform for archiving.

[0055] If the FI value is still 0.25 after verification, the system determines that "rinsing is unqualified". The vehicle and the back-end system issue a warning reminder, and the vehicle is returned for a second rinse, or a second rinse is performed at a fixed point on the vehicle's return trip. During the second rinse, the system automatically raises the risk level to "Level IV High Risk" and generates a stronger set of rinsing instructions, such as: {water pressure: 1.8 MPa, continuous water jet time: 10 seconds, repeated horizontal and vertical vibration}, then performs the second rinse and re-verifies until it passes.

[0056] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart water flushing method based on the degree of dirtiness of insulators, characterized in that, Includes the following steps: Step 1: Insulator identification, positioning, and tracking; The water washing truck is equipped with vision cameras, lidar and Beidou positioning at the front and rear ends. Through multi-sensor fusion technology, it identifies and locates insulators and pole numbers, thereby controlling and guiding the automated water cannon to track and aim at the insulator target in real time, ready to wash the insulators at any time. Step 2: Generate an insulator database; After identifying the insulators and pole numbers, the coordinates of the line map are drawn using the BeiDou positioning system. Then, the corresponding pole numbers and insulator numbers identified and located are input into the line map, thus forming a line network database. Step 3: Real-time data collection of dirt and grime on the insulator surface; By using an insulator surface dirt detection device mounted on a water washing vehicle, the surface condition parameters and leakage current parameters of the target insulator can be obtained in real time in a non-contact manner. Step 4: Dynamically assess the risk of insulator leakage current; The surface condition parameters and leakage current parameters are processed, and the comprehensive contamination index of the insulator is calculated through the contamination assessment model. The leakage current parameters are then integrated to dynamically assess the current risk level of the insulator. Step 5: Generate an insulator flushing strategy; Based on the current risk level and the shape and size of the insulator, a corresponding flushing control command is generated by matching from a predefined flushing strategy library. The flushing control command at least defines the flushing water pressure, flushing duration, and the trajectory of the flushing jet sweeping the insulator. Step 6: Perform a precise rinse; The water flushing system performs automated flushing operations on the target insulator according to the flushing control command.

2. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 1, characterized in that, It also includes the following steps: Step 7: Verify the rinsing effect; After the washing operation is completed, steps 3 and 4 are performed again using the insulator surface dirt detection equipment at the other end of the vehicle to obtain the status parameters of the insulator after washing and assess its risk level. If the overall dirt index after rinsing is below the safety threshold and the risk level drops to low, the rinsing is deemed qualified; otherwise, the rinsing is deemed unqualified, and the insulator data is listed as a warning. The rinsing strategy is then optimized based on the data after rinsing, and the process awaits a second rinsing.

3. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 1, characterized in that, In step 3, the surface state parameters are obtained through optical sensing or image recognition technology, including estimated values ​​of equivalent salt density (ESDD) and equivalent gray density (NSDD). The leakage current parameters are obtained using ultrasonic technology, including the effective value of the leakage current and the pulse frequency.

4. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 3, characterized in that, In step 4, the dirt assessment model is a model trained based on machine learning algorithms. Its input features include ESDD, NSDD, ambient humidity and temperature, and its output is a comprehensive dirt index. The specific process of dynamically assessing the risk level is as follows: the comprehensive dirt index is weighted and fused with the effective value of the leakage current and the pulse frequency, and the risk level is divided into five levels: high, medium, low and low, based on the preset threshold range.

5. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 4, characterized in that, In step 5, the predefined flushing strategy library contains flushing strategies for different risk levels as follows: For higher and higher risk levels, an instruction is generated to use a powerful flushing mode with high water pressure, large water volume, and long duration. For medium-risk levels, an instruction is generated to use a standard flushing mode with medium water pressure and volume. For low-risk levels, an instruction is generated to use an energy-saving flushing mode with low water pressure and small water volume; For lower risk levels, generate an instruction to temporarily not flush but only record the status; Meanwhile, different sweeping trajectories are used to clean insulators of different shapes and sizes: horizontal bracket insulators are cleaned by left and right sweeping and up and down vibration; oblique bracket insulators are cleaned by stepped up and down sweeping; and return wire insulators are cleaned by left and right swinging.

6. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 1, characterized in that, It also includes historical data learning: collecting and learning historical flushing data, environmental data, and insulator fault data to dynamically optimize the parameters of the dirt assessment model and the flushing strategies in the predefined flushing strategy library.

7. The intelligent water flushing method based on the degree of dirtiness of insulators according to claim 2, characterized in that, It also includes uploading all data generated in steps 1-7, including detection data, risk level, flushing parameters and verification results, to the remote management platform to form a digital archive of the insulator's condition and maintenance throughout its entire lifecycle, for subsequent analysis and decision support.

Images