Ship lock metal structure detection system and method based on high-precision grating strain gauge technology

Abstract

A ship lock metal structure detection system and method based on a high-precision grating strain gauge technology comprises a shell and a plurality of first oil cylinders, a plurality of valve blocks and a plurality of second oil cylinders are fixedly arranged in the shell, stepped holes are formed in the valve blocks, shaft tubes are assembled in slim holes in the stepped holes, and the ends, located outside the slim holes, of the shaft tubes are fixedly connected with a jacking frame. A clamping sleeve is in threaded connection with one end of a thick hole in the stepped hole, a first spring sleeves the shaft tube, two ends of the first spring abut against the valve block and the jacking frame respectively, a piston rod of the second oil cylinder is fixedly connected with the jacking frame, a grating strain gauge is arranged in the stepped hole, and two ends of the grating strain gauge are fixedly connected with the clamping sleeve and the shaft tube through clamping pieces respectively; the two ends of the first oil cylinder are connected to the hinged supports respectively, and the first oil cylinder is communicated with the second oil cylinder through an oil pipe. The method is used for solving the problem that when the miter gate is in a running state, manual detection cannot be completed, and consequently the comprehensive detection result is inaccurate.

Description

Technical Field

[0001] The present invention relates to a ship lock metal structure detection system and method based on high-precision grating strain gauge technology. Background Art

[0002] In the field of hydraulic engineering, gates, as key facilities for controlling water discharge (release) channels, play a vital role in regulating water levels and controlling water flows. Miter gates, due to their unique structural advantages, have become a widely used hydraulic device in navigable river locks. However, miter gates face numerous challenges in their actual operation. For one thing, their gate leaves are large in mass and size, representing a typical ultra-large, thin-walled structural component with an opening. Furthermore, during frequent opening and closing operations, miter gates continuously withstand complex external forces such as dynamic water loads and water flow impacts.

[0003] These unfavorable factors lead to poor operating conditions and extremely complex stress conditions for the miter gate, which in turn makes its structure prone to cracks. Once cracks appear in the miter gate, it will inevitably have a negative impact on its overall strength, posing a serious safety hazard.

[0004] Currently, crack detection for miter gates relies primarily on manual labor. While manual inspection is effective when the gates are stationary, however, once the gates are operational, the turbulent water flow and hazardous working environment make it impossible for humans to access and complete the inspection task. This makes it impossible to obtain real-time and accurate crack information during the operation of the gates, ultimately leading to biased comprehensive inspection results. This makes it difficult to truly reflect the actual condition of the gates, and thus fails to provide a reliable basis for the safe maintenance of water conservancy facilities. Summary of the Invention

[0005] The purpose of the present invention is to provide a ship lock metal structure detection system and method based on high-precision grating strain gauge technology, which is used to solve the problem that when the miter gate is in operation, manual detection cannot be completed, resulting in inaccurate comprehensive detection results.

[0006] In order to solve the above problems, the technical solution of the present invention is: A ship lock metal structure detection system based on high-precision grating strain gauge technology includes a shell and multiple first oil cylinders, multiple valve blocks and multiple second oil cylinders are fixedly arranged in the shell, a stepped hole is provided in the valve block, a shaft tube is assembled in the fine hole in the stepped hole, the shaft tube is located outside the fine hole and is fixedly connected to the top pressure frame, a ferrule is threadedly connected to one end of the coarse hole in the stepped hole, a first spring is arranged on the outer sleeve of the shaft tube, the two ends of the first spring respectively rest on the valve block and the top pressure frame, the second oil cylinder piston rod is fixedly connected to the top pressure frame, a grating strain gauge is provided in the stepped hole, the two ends of the grating strain gauge are respectively fixedly connected to the ferrule and the shaft tube by clamping parts; the two ends of the first oil cylinder are respectively connected to a hinge seat, the first oil cylinder is connected to the second oil cylinder through an oil pipe, and when the same hydraulic oil is input to the first oil cylinder and the second oil cylinder, the piston rod stroke of the first oil cylinder is smaller than the piston stroke of the second oil cylinder.

[0007] A collecting pipe is arranged in the shell. The collecting pipe has an oil discharge port and a plurality of oil inlets. Each first oil cylinder is connected with each oil inlet through an oil pipe, and the oil discharge port is connected with the second oil cylinder through an oil pipe.

[0008] The shell includes a shell and a shell cover, the shell and the shell cover are sealed together, a plurality of through-wall straight-through pipe joints are installed on the shell, a seal is provided between the through-wall straight-through pipe joints and the shell, the oil pipes inside and outside the shell are connected to both ends of the through-wall straight-through pipe joints, a plurality of waterproof joints are installed on the shell, and the optical fiber connected to the grating strain gauge is assembled in the waterproof joint.

[0009] The first oil cylinder is a spring return oil cylinder.

[0010] The clamping part includes a mounting hole opened at one end of the ferrule and the shaft tube, and the mounting hole includes a countersunk hole, a tapered hole and a threaded hole. A second spring is provided in the countersunk hole, and a plurality of fan-shaped tiles are provided in the tapered hole. The multiple fan-shaped tiles are matched with the tapered hole in an inclined surface. The threaded hole is threadedly connected with a clamping sleeve, and the clamping sleeve pushes the fan-shaped tiles against the second spring.

[0011] It also includes a plurality of pipe clamps, which include clamping blocks. The bottom surface of the clamping block is provided with a clamping groove for fixing the oil pipe, and magnets are fixedly connected to the clamping block on both sides of the clamping groove.

[0012] It also includes a signal transmitter, a multi-channel data conversion module, a data demodulator and a data analysis and processing module. Each grating strain gauge transmits the collected signal to the multi-channel data conversion module after passing through the signal transmitter. The multi-channel data conversion module transmits the signal to the data demodulator. The data demodulator converts the analog signal into a digital signal and sends it to the data analysis and processing module. After the data analysis and processing module processes the data, the data is displayed on the display.

[0013] The data analysis and processing module has a built-in crack extension prediction algorithm. The algorithm is based on time series analysis and machine learning models. It generates a crack development trend prediction curve through historical data training and triggers an alarm when the prediction curve exceeds a safety threshold.

[0014] The prediction algorithm comprises the following steps: a) performing sliding average filtering on displacement data collected by the grating strain gauge (14); b) Using LSTM neural network to establish a dynamic model of crack propagation; c) Conduct multivariate regression analysis based on parameters such as ambient temperature and water pressure; d) Update model parameters through Bayesian optimization algorithm.

[0015] The beneficial effects of the present invention are: 1. High-Precision Detection: This invention utilizes high-precision grating strain gauges to detect minute crack changes with an accuracy of up to 0.01mm. This feature enables the system to accurately monitor the development of cracks in miter gates, providing strong support for the timely identification of potential safety hazards.

[0016] 2. Amplification of Crack Changes: By cleverly designing the piston diameters of the first and second cylinders, the pistons produce a difference in stroke when the same hydraulic oil is input, effectively amplifying the crack change. This design significantly improves the sensitivity of the detection system, allowing even subtle crack changes to be clearly detected.

[0017] 3. Real-time Monitoring and Display: The system is equipped with a signal transmitter, a multi-channel data conversion module, a data demodulator, and a data analysis and processing module, all working together. The signals collected by the grating strain gauges undergo a series of processing steps, quickly displaying the crack status on the display in real time. This allows staff to promptly monitor the crack status of the miter gate and promptly implement appropriate maintenance measures.

[0018] 4. Reduced costs: When multiple cracks exist in the same door grille, multiple primary cylinders can share a single secondary cylinder and grating strain gauge. This design significantly reduces the number of key components used in the detection system, effectively lowering overall costs.

[0019] 5. Improved system stability and reliability: The manifold allows for centralized management and rational distribution of hydraulic oil, ensuring more stable and reliable system operation. Furthermore, the housing's waterproof design, including the use of gaskets, sealing rings, and waterproof connectors, ensures the system can operate normally in harsh underwater environments, unaffected by water pressure and humidity, further enhancing system reliability.

[0020] 6. Accurately locate excessive cracks: When multiple first cylinders aggregate the crack changes to the second cylinder and feed them back to the grating strain gauge, if the total change exceeds the preset value, the staff can quickly lock the door grid where the excessive cracks are located, and then conduct targeted inspections on each crack individually, greatly improving detection efficiency.

[0021] 7. Eliminate mechanical backlash: A spring-return cylinder is used as the first cylinder, effectively eliminating the backlash between the cylinder and the hinge. This not only improves the accuracy of the detection system, but also increases the system's response speed, reduces measurement errors caused by mechanical vibration or impact, and ensures that crack changes can be accurately detected.

[0022] 8. Optimized clamping structure: The clamping element uses multiple fan-shaped tiles to cooperate with the inclined surface of the tapered hole to provide uniform and stable clamping force, firmly fixing the grating strain gauge. At the same time, the second spring acts as a buffer during the clamping process, preventing excessive clamping force from damaging the grating strain gauge, and comprehensively protecting the sensing element.

[0023] 9. Easy Installation and Maintenance: The magnetic pipe clamp design effectively prevents tubing from becoming entangled during installation, making the entire inspection device layout more neat and aesthetically pleasing. Furthermore, the magnetic clamp's fastening facilitates installation and removal, significantly improving on-site installation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present invention will be further described below with reference to the accompanying drawings: Figure 1 is a schematic diagram of the three-dimensional structure of the present invention, Figure 2 This is a schematic diagram of the main structure of the present invention, Figure 3 It is a schematic diagram of the cross-sectional structure of the present invention, Figure 4 for Figure 3 Schematic diagram of the local enlarged structure at A in the middle, Figure 5 for Figure 4 Schematic diagram of the cross-sectional structure at BB in the middle, Figure 6 This is a schematic diagram of the structure of the pipe clamp of the present invention. Figure 7 This is a connection diagram of the functional modules of the system of the present invention.

[0025] In the figure: miter gate 1, housing 2, first oil cylinder 3, hinge seat 4, pipe clamp 5, oil pipe 6, clamping part 7, ferrule 8, shaft tube 9, top pressure frame 10, second oil cylinder 11, collecting pipe 12, through-wall straight pipe 13, grating strain gauge 14, waterproof joint 15, valve block 16, first spring 17, signal transmitter 18, multi-channel data conversion module 19, data demodulator 20, data analysis and processing module 21, display 22, magnet 51, card block 52, compression sleeve 71, sector tile 72, second spring 73. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] Example 1: like Figures 1 to 3 As shown, a ship lock metal structure detection system based on high-precision grating strain gauge technology includes a housing 2 and multiple first oil cylinders 3. Multiple valve blocks 16 and multiple second oil cylinders 11 are fixedly provided in the housing 2. A stepped hole is provided in the valve block 16. A shaft tube 9 is installed in the fine hole in the stepped hole. The shaft tube 9 can move axially in the fine hole. The shaft tube 9 is located outside the fine hole and is fixedly connected to the top pressure frame 10. A ferrule 8 is threadedly connected to one end of the coarse hole in the stepped hole. A first spring 17 is provided on the outer sleeve of the shaft tube 9. The two ends of the first spring 17 respectively abut against the valve block 16 and the top pressure frame 10, the piston rod of the second oil cylinder 11 is fixedly connected to the top pressure frame 10, and an FGB grating strain gauge 14 is provided in the stepped hole. The two ends of the grating strain gauge 14 are fixedly connected to the ferrule 8 and the shaft tube 9 through the clamping piece 7; the two ends of the first oil cylinder 3 are respectively connected to a hinge seat 4, and the first oil cylinder 3 is connected to the second oil cylinder 11 through the oil pipe 6, and the piston diameter of the first oil cylinder 3 is larger than that of the second oil cylinder 11. In this way, when the same hydraulic oil is input into the first oil cylinder 3 and the second oil cylinder 11, the piston rod stroke of the first oil cylinder 3 is smaller than the piston stroke of the second oil cylinder 11.

[0028] like Figure 7 As shown, it also includes a display, a signal transmitter, a multi-channel data conversion module, a data demodulator and a data analysis and processing module. Each grating strain gauge 14 transmits the collected signal to the multi-channel data conversion module after passing through the signal transmitter. The multi-channel data conversion module transmits the signal to the data demodulator. The data demodulator converts the analog signal into a digital signal and sends it to the data analysis and processing module. After the data analysis and processing module processes the data, it displays the data through the display. The data analysis and processing module is an embedded computer.

[0029] Model of each component: Grating strain gauge 14: Model FGB-1000, range ±10mm, accuracy 0.01mm; Signal transmitter 18: Model ST-200, input range 0-10V, output range 0-5V; Multi-channel data conversion module 19: Model DMC-300; Data demodulator 20: Model DMD-50, conversion accuracy 16 bits; Data analysis and processing module 21: Model ADP-800, embedded computer, processing speed 1 GHz; Display 22: Model LCD-15, resolution 1024×768; The installation method includes the following steps: Initially, the first oil cylinder 3 and the second oil cylinder 11 are filled with hydraulic oil; Step 1: Secure the hinges: Open the bleed screws on the first cylinder 3 and secure the hinges 4 at both ends of the first cylinder 3 to the door on both sides of the crack using either casting adhesive or welding. Next, bond the outer shell 2 to the door using casting adhesive. Adjust the piston in the second cylinder 11 to retract to the center position, and finally tighten the bleed screws on the first cylinder 3.

[0030] Step 2. Install the grating strain gauge: Connect and fix one end of the grating strain gauge 14 to the clamping piece 7 at one end of the shaft tube 9, then put the first spring 17 and the valve block 16 on the outer sleeve of the shaft tube 9, squeeze the first spring 17 at the same time, so that the first spring 17 is in a compressed state, and use screws to fix the valve block 16 to the housing 2, then thread the ferrule 8 to one end of the stepped hole coarse hole, and rotate the ferrule 8 to fine-tune the position of the clamping piece 7 on the ferrule 8, so that the clamping piece 7 on the ferrule 8 is moved to the installation position of the other end of the grating strain gauge 14, and finally tighten the clamping piece 7 to fix the other end of the grating strain gauge 14; Step 3, adjust the spring state: slowly loosen the bleed screw on the first oil cylinder 3, slowly move the valve block 16, let the first spring 17 press the piston rod of the second oil cylinder 11, so that the hydraulic oil in the first oil cylinder 3 drips out from the bleed screw, then immediately tighten the bleed screw, while requiring the first spring 17 to remain in a compressed state, so as to ensure that the grating strain gauge 14 is in a tensile state, finally use the bolts to connect the valve block 16 to the housing 2; Step 4: Adjust the spring state: Connect the optical fibers at both ends of the grating strain gauge 14 to the signal transmitter; the data analysis and processing module resets the signal transmitted by the grating strain gauge 14 collected at this time to zero, thus completing the installation process of the detection device.

[0031] The working process of the present invention is as follows: when the miter gate is in operation, the crack expands and contracts, and the first cylinder 3 follows the movement of the crack. Since the first cylinder 3 and the second cylinder 11 are connected by the oil pipe 6, and the piston diameter of the first cylinder 3 is larger than that of the second cylinder 11, under the same hydraulic oil input, the piston rod stroke of the first cylinder 3 is smaller than the piston rod stroke of the second cylinder 11, thereby amplifying the crack change. The movement of the piston rod of the second cylinder 11 drives the top pressure frame 10 to move, which in turn drives the grating strain gauge 14 to expand and contract, causing its wavelength to change. The grating strain gauge 14 sends the wavelength change signal to the signal transmitter 18. After the signal is processed by the signal transmitter 18, it is transmitted to the data demodulator 20 through the multi-channel data conversion module 19. The data demodulator 20 converts the analog signal into a digital signal and sends it to the data analysis and processing module 21. The data analysis and processing module 21 processes and stores the data and displays the crack status on the display 22.

[0032] Technical Effect: The high-precision detection of the grating strain gauge 14 enables real-time monitoring of minute crack changes with an accuracy of up to 0.01mm. The differential stroke design between the first and second cylinders 3 and 11 effectively amplifies crack changes and improves detection sensitivity. The embedded computer's data analysis and processing module 21 rapidly processes signals and displays crack status in real time, facilitating the timely identification of safety hazards.

[0033] Example 2: like Figure 3 As shown, a manifold 12 is provided in the housing 2 , and the manifold 12 has an oil discharge port and multiple oil inlets. Each first oil cylinder 3 is connected to each oil inlet through an oil pipe 6 , and the oil discharge port is connected to the second oil cylinder 11 through the oil pipe 6 .

[0034] Working Principle: When multiple cracks exist within the same gate section of a miter gate, the first cylinder 3 located at each crack will independently operate based on the crack's variation. Because multiple first cylinders 3 share a single second cylinder 11 and grating strain gauge 14, the crack variation is aggregated by these first cylinders 3 and fed to the second cylinder 11. The second cylinder 11 then amplifies this aggregated variation and feeds it back to the grating strain gauge 14. If the total variation detected by the grating strain gauge 14 exceeds a preset value, it indicates that at least one crack in that gate section has exceeded the specified value. At this point, personnel can individually inspect each crack in the gate section to pinpoint the location of the crack that exceeds the specified value.

[0035] Technical Effect: This reduces the use of grating strain gauges 14 and second oil cylinders 11, lowering the cost of the detection system. Centralized management of hydraulic oil distribution improves system stability and reliability. It also enables rapid location of the door grid where excessive cracks are located, improving detection efficiency.

[0036] Example 3: The housing 2 includes a shell and a shell cover, a sealing gasket is provided between the shell and the shell cover, a plurality of through-wall straight-through pipe 13 joints are installed on the shell, a sealing member is provided between the through-wall straight-through pipe 13 joint and the shell, the sealing member is a sealing ring, the oil pipes 6 inside and outside the shell are connected to the two ends of the through-wall straight-through pipe 13 joint, a plurality of waterproof joints 15 are installed on the shell, and the optical fiber connected to the grating strain gauge 14 is assembled in the waterproof joint 15.

[0037] Working principle: In an underwater environment, the shell and shell cover of the housing 2 are sealed by a gasket and a sealing ring to prevent water from entering the interior of the housing. The use of the through-wall straight pipe 13 connector and the waterproof connector 15 ensures that the oil pipe 6 and the optical fiber will not leak when passing through the housing. The grating strain gauge 14 is installed in a sealed space and can work normally even underwater. When the crack changes, the operating principle of the first oil cylinder 3 and the second oil cylinder 11 is the same as that of Example 1. The change signal detected by the grating strain gauge 14 is transmitted to the signal processing module through the optical fiber in the waterproof connector 15, and finally the crack status is displayed on the display 22.

[0038] Technical Effect: The waterproof design ensures the detection system can operate properly in underwater environments, unaffected by water pressure and humidity. The dual sealing measures of the gasket and sealing ring improve the system's waterproof performance and reliability. The use of waterproof connector 15 protects the optical fiber connection from water immersion, ensuring stable signal transmission.

[0039] Example 4: like Figure 2 As shown, the first oil cylinder 3 is a spring return oil cylinder. This structure can eliminate the gap between the first oil cylinder 3 and the hinge seat 4, thereby improving the detection accuracy.

[0040] Working Principle: When the spring-return cylinder is activated by hydraulic fluid, the spring provides a reverse force, allowing the cylinder to automatically return to its initial position when no external force is applied. As the crack changes, the first cylinder (3) follows the crack. The spring-return cylinder design eliminates mechanical backlash, ensuring accurate detection of crack changes. Furthermore, the spring's reverse force reduces measurement errors caused by mechanical vibration or shock, improving detection accuracy.

[0041] Technical Effect: The spring return cylinder design eliminates mechanical backlash, improving the accuracy and response speed of the detection system. It also reduces measurement errors caused by mechanical backlash and ensures that crack changes can be accurately detected.

[0042] Example 5: like Figure 4 and 5As shown, the clamping part 7 includes a mounting hole opened at one end of the sleeve 8 and the shaft tube 9, and the mounting hole includes a countersunk hole, a tapered hole and a threaded hole. A second spring 73 is provided in the countersunk hole, and a plurality of fan-shaped tiles 72 are provided in the tapered hole. The plurality of fan-shaped tiles 72 are matched with the tapered hole at an inclined surface, and a clamping sleeve 71 is threadedly connected to the threaded hole, and the clamping sleeve 71 pushes the fan-shaped tiles 72 against the second spring 73.

[0043] Working Principle: The design of the clamping element 7 utilizes multiple sector-shaped tiles 72 that mate with the tapered surface of the tapered hole to provide uniform clamping force. When the compression sleeve 71 is tightened, the sector-shaped tiles 72 are pushed toward the second spring 73, clamping the strain gauge 14 securely. The second spring 73 provides a buffer during the clamping process, preventing damage to the strain gauge 14 caused by excessive clamping force. This design not only ensures a secure hold for the strain gauge 14 but also provides a degree of protection from mechanical shock.

[0044] Technical Effect: The optimized design of the clamping member 7 ensures a secure fixation of the strain gauge 14, preventing measurement errors caused by loosening during testing. The multiple sector-shaped tiles 72, in conjunction with the inclined surface of the tapered hole, provide uniform clamping force, protecting the strain gauge 14 from damage. The use of the second spring 73 provides a cushion during the clamping process, further protecting the strain gauge 14.

[0045] Example 6: like Figure 6 As shown, it also includes a plurality of pipe clamps 5, each of which includes a clamping block 52. The bottom surface of the clamping block 52 is provided with a clamping groove for fixing the oil pipe 6, and NdFeB magnets 51 are fixedly connected to the clamping block on both sides of the clamping groove.

[0046] Working Principle: The pipe clamp 5 is attached to the metal surface of the miter gate via a neodymium iron boron magnet 51. The slots in the clamping block 52 secure the oil pipe 6. This design effectively prevents the oil pipe 6 from becoming entangled during installation and enhances the overall appearance of the inspection device. The magnetically secured pipe clamp 5 facilitates installation and removal, improving on-site installation and maintenance efficiency.

[0047] Technical Effect: The magnetically secured pipe clamp 5 effectively prevents entanglement of the oil pipes 6, improving system reliability. This provides a neat and aesthetically pleasing appearance for the crack detection device mounted on the miter gate, facilitating on-site installation and maintenance. The magnetic securing of the pipe clamp 5 facilitates installation and removal, improving on-site operation efficiency.

[0048] In actual application, this detection system was installed on the miter gate of a large ship lock. The miter gate is 10 meters high, 6 meters wide, and 1 meter thick. Multiple mounting points were pre-set on the gate body to secure the hinge seat 4 and the housing 2. During installation, the above steps were strictly followed to ensure the accuracy of each component.

[0049] After installation, the system was debugged. The system's response was tested by simulating crack changes. A micrometer was used to measure the actual crack change, while simultaneously recording the data from the grating strain gauge 14 and comparing the two for consistency. After multiple tests, the system's detection accuracy was stable within 0.01 mm, and its response speed was fast, enabling it to reflect crack changes in real time.

[0050] Example 7: During actual operation, the lock undergoes multiple opening and closing operations daily, with turbulent water flows and harsh environments. The detection system operates stably in this environment, displaying the crack status in real time on the display 22. Staff can review the data at any time and promptly identify potential safety hazards. During one operation, the system detected that the change in a crack at a certain location exceeded a preset value. Staff immediately conducted a detailed inspection of the area and discovered that the crack was expanding. Prompt repair measures were taken, averting a potential accident.

[0051] To further verify the reliability and accuracy of this detection system, we compared it with traditional detection methods. We selected the same area of the same miter gate and monitored it using both manual inspection and this system. Manual inspection was performed while the gate was stationary, using high-precision measurement tools to measure crack size; the system continuously monitored the gate while it was in operation.

[0052] After a period of comparison, it was discovered that manual monitoring only captures static data and, due to environmental constraints, cannot be performed while the gate is in operation. This system, however, provides continuous, real-time monitoring data with higher accuracy. When it comes to detecting crack changes, this system can proactively detect subtle trends, providing strong support for preventive maintenance. Furthermore, this system provides comprehensive monitoring of key areas, and the data analysis and processing module 21 conducts in-depth analysis of the collected data to generate detailed crack development reports.

[0053] Based on the report, staff can accurately determine which cracks require priority treatment and rationally allocate maintenance resources. Furthermore, by predicting crack development trends and formulating maintenance plans, the service life of the locks has been effectively extended and maintenance costs have been reduced.

[0054] To accommodate cracks of varying sizes and shapes, the detection system has been modified. Clamping elements 7 of various sizes have been designed to accommodate grating strain gauges 14 of varying diameters. Furthermore, the piston diameter ratio between the first and second cylinders 3 and 11 has been adjusted to accommodate the amplification requirements for varying crack variations.

[0055] In some cases, such as when cracks are located in hard-to-reach locations, the detection system can be extended to cover a wider area by extending the length of the oil pipe 6. Furthermore, a wireless data transmission module has been developed, allowing monitoring data to be transmitted to a remote monitoring center via a wireless network, facilitating real-time monitoring and analysis at various locations.

[0056] The data analysis and processing module processes and stores the data and predicts the crack development trend. The prediction method includes the following steps: S1. Data Collection and Preprocessing First, the crack width y detected by the grating strain gauge at different times is collected. i and length data, as well as the corresponding environmental conditions (such as temperature T i , humidity Hi), lock operation status (such as opening and closing frequency f i , water flow velocity v i The data collection time interval can be set according to actual needs, such as once every hour or every day.

[0057] The collected data is cleaned to remove outliers and noise. For example, the 3δ principle can be used, which removes data points that are outside the mean ±3 times the standard deviation. The data is then normalized to scale all data to the (0,1) interval using the following formula:

[0058] Among them, x is the original data, x min and x max The minimum and maximum values of the data column respectively.

[0059] S2. Establish a crack development trend model S2.1 Linear regression model Assuming that the crack width y changes linearly with time t, a linear regression model can be established: y=a+bt

[0060] Where a is the intercept and b is the slope, which represent the crack growth rate. The parameters a and b are estimated by the least squares method:

[0061]

[0062] Where ti and yi are the time and crack width of the i-th measurement, respectively. and are the average values of time t and crack width y, respectively.

[0063] Using the above formula to calculate a and b, we can predict the future time point t future The crack width yfuture : y future =a+bt future

[0064] S2.2 Time Series Analysis Model The crack width data is considered as a time series {y t}, considering its own lag term and moving average term, establish the ARIMA(p,d,q) model:

[0065] Among them, B is the lag operator, φ i is the autoregressive coefficient, θ j is the moving average coefficient, d is the difference order, is the white noise error term.

[0066] The appropriate values of p, d, and q are determined by ACF (autocorrelation function) and PACF (partial autocorrelation function) graphs. After fitting the ARIMA model with historical data and obtaining the model parameters, the model can be used to predict future values. For example, predict the crack width y in the next k steps. t +k.

[0067] S2.3 Machine Learning Model Use neural networks to predict crack development trends. Assume the input layer has n neurons (such as the current crack width, length, ambient temperature, humidity, number of lock operations, etc.), the hidden layer has m neurons, and the output layer predicts the future crack width. The forward propagation formula of the neural network is:

[0068] Among them, x j is the input feature, is the weight from the input layer to the hidden layer, is the bias of the hidden layer, is the weight from the hidden layer to the output layer, is the bias of the output layer, and δ is the activation function (such as ReLU or Sigmoid).

[0069] The network parameters are adjusted by backpropagation algorithm and gradient descent method to minimize the mean square error (MSE) between the predicted value and the actual value:

[0070] Where yi is the actual crack width, is the predicted value, and N is the number of samples.

[0071] After training is completed, the neural network model is used to predict the crack width of new input data.

[0072] S2.4 Multifactor association model Considering the linear relationship between crack width y and factors such as temperature T, humidity H, and opening and closing frequency f, a multiple linear regression model is established: y = a + b1T + b2H + b3f Estimate model parameters a, b1, b2, b3, etc. by the least squares method:

[0073] in, is the parameter vector, X is the design matrix containing the values of all independent variables, and y is the dependent variable vector.

[0074] Using the estimated parameters, the crack width variation under given environmental and operating conditions can be predicted.

[0075] S3. Model training and optimization The collected historical data is divided into a training set and a test set, and the established model is trained using the training set. For linear regression and multivariate linear regression models, the parameters are estimated using the least squares method; for ARIMA models, the model parameters are determined using the maximum likelihood estimation method; for neural network models, the network weights are adjusted using backpropagation and gradient descent.

[0076] Validate and evaluate the trained model on the test set, and calculate the error indicators between the predicted values and the actual values, such as the mean square error (MSE):

[0077] Mean Absolute Error (MAE):

[0078] The model is optimized and adjusted based on the evaluation results, such as adjusting the model structure, adding or reducing features, etc., to improve the model's prediction accuracy and generalization ability.

[0079] S4. Consider the influence of multiple factors Analyze and determine various factors that affect the development trend of cracks, such as environmental factors (temperature, humidity, water quality, etc.), lock operating conditions (opening and closing frequency, water flow impact force, water pressure, etc.), material properties (material, strength, fatigue performance of metal structures, etc.), and the geometric shape and expansion direction of the cracks themselves.

[0080] These factors can be incorporated into the prediction model as input variables, or a multi-factor association model can be established to analyze the quantitative relationship between each factor and crack development trends. For example, a multivariate linear regression model can be established with factors such as temperature T, humidity H, and opening and closing frequency f as independent variables, and the crack width change rate Δy / Δt as the dependent variable. Through regression analysis, the weight coefficient of each factor can be determined to predict the development trend of cracks under different environmental and operating conditions.

[0081] S5. Real-time updates and feedback During the operation of the lock metal structure inspection system, new crack detection data is continuously collected and compared with the model prediction results. Based on the deviation between the actual detection data and the predicted data, the parameters or structure of the prediction model are adjusted and updated in a timely manner. Therefore, it is necessary to update the model parameters in real time to ensure the accuracy and reliability of the prediction results. The following are three real-time update methods: S5.1. Incremental Learning Methods Incremental learning is a method for dynamically updating model parameters, suitable for scenarios where data is constantly updated. It adjusts model parameters by gradually introducing new data without retraining the entire model.

[0082] S5.1.1 Incremental Least Squares Suppose a linear regression model y = a + bt is used to predict the change in crack width y over time t. The initial model parameters a and b can be calculated using the least squares method. When a new data point (tnew, ynew) arrives, the parameters can be updated using the incremental least squares method.

[0083] The core idea of the incremental least squares method is to use new data points to modify the existing model instead of recalculating the entire model. The specific steps are as follows: Compute the residual error of the model for a new data point: residual=y new -(a+bt new )

[0084] Adjust model parameters based on residuals: Δa=learning_rate×residual

[0085] Δb=learning_rate×residual×t new

[0086] Update model parameters: a new =a+Δa b new =b+Δb

[0087] Δa represents the adjustment amount of the model intercept a, which is used to adjust the intercept of the model to better fit the new data point; Δb represents the adjustment amount of the model slope b, which is used to adjust the slope of the model, residual represents the residual of the new data point, y new Represents the newly measured crack width, which is used to calculate the residual of the model and adjust the model parameters according to the residual to improve the prediction accuracy of the model. new Is the new data point time, fitting the new data point learning_rete is the learning rate, used to control the step size of parameter update.

[0088] S5.1.2 Sliding Window Method The sliding window method updates model parameters in real time by maintaining a fixed-size data window. When new data arrives, the oldest data point in the window is removed and the new data is added to the window. The model is then retrained using the data within the window.

[0089] For example, assuming the window size is N, the data in the current window is {(t1, y1), (t2, y2), ..., (t N ,y N )}. When a new data point (tN+1, YN+1) arrives, the oldest data point (t1, y1) is removed and the window is updated to {(t2, y2), (t3, y3), ..., (t N+1 ,y N+1 )}. The model parameters a and b are then recalculated using the data within the window.

[0090] S5.2 Online Learning Algorithm Online learning algorithms are a method for dynamically adjusting model parameters, suitable for scenarios where data streams are constantly updated. They adjust model parameters by gradually introducing new data without retraining the entire model.

[0091] S5.2.1 Stochastic Gradient Descent (SGD) Stochastic gradient descent is an online learning algorithm that is suitable for large datasets. It updates the model parameters by processing data points one by one instead of using the entire dataset.

[0092] Suppose we use a linear regression model y = a + bt to predict the change in crack width y over time t. The initial model parameters a and b can be calculated using the least squares method. When a new data point (tnew, ynew) arrives, the parameters can be updated using stochastic gradient descent.

[0093] The core idea of stochastic gradient descent is to use the gradient information of new data points to adjust the model parameters. The specific steps are as follows: Calculate the gradient of the new data point:

[0094]

[0095] Adjust model parameters based on gradients:

[0096]

[0097] Among them, learning_rete is the learning rate, which is used to control the step size of parameter update.

[0098] S5.3. Adaptive Filtering Methods Adaptive filtering methods update model parameters in real time by dynamically adjusting filter parameters. They are suitable for signal processing and dynamic system modeling.

[0099] S5.3.1 Least Mean Squares (LMS) filter The minimum mean square error filter is an adaptive filter that adjusts filter parameters by minimizing the square of the prediction error. It is suitable for real-time signal processing and dynamic system modeling.

[0100] Suppose we use a linear regression model y = a + bt to predict the change in crack width y over time t. The initial model parameters a and b can be calculated using the least squares method. When a new data point (tnew, ynew) arrives, the parameters can be updated using a minimum mean square error filter.

[0101] The core idea of the minimum mean square error filter is to use the error information of the new data point to adjust the filter parameters. The specific steps are as follows: Compute the error for a new data point: e new =y new -(a+bt new )

[0102] Adjust the filter parameters according to the error: a new =a+μe new

[0103] b new =b+μe new t new

[0104] Among them, μ is the step size parameter, which is used to control the step size of parameter update, e new Represents the prediction error or residual of a new data point, which is used to evaluate the deviation between the model prediction value and the actual measurement value and serves as the basis for adjusting the model parameters.

[0105] Real-time updating of model parameters ensures accurate and reliable predictions of crack development trends. Incremental learning methods, online learning algorithms, and adaptive filtering techniques can dynamically adjust model parameters to adapt to the dynamic characteristics of crack changes. These methods not only improve prediction accuracy but also reduce computing resource consumption, making them suitable for real-time monitoring and early warning systems.

[0106] It can be seen that crack prediction has the following technical effects: 1. Improve prediction accuracy Real-time data updates: By collecting crack detection data in real time and dynamically updating model parameters, the latest trends in crack changes can be reflected promptly, avoiding prediction bias caused by data lag. For example, if the crack growth rate suddenly accelerates, the model can quickly capture this change and adjust the prediction results, thereby more accurately predicting the future crack growth.

[0107] Comprehensive application of multiple models: Combining linear regression, time series analysis, and machine learning models leverages the strengths of each model. Linear regression models are simple and efficient, suitable for situations where crack changes are relatively linear. Time series analysis models can account for the temporal correlation of crack changes and are suitable for situations where crack expansion exhibits periodicity or trends. Machine learning models can handle complex nonlinear relationships, further improving prediction accuracy.

[0108] Consider the influence of multiple factors: Incorporate multiple factors such as temperature, humidity, and lock operation status into the prediction model, and comprehensively consider the impact of these factors on the development trend of cracks, making the prediction results more accurate and reliable.

[0109] 2. Enhance model adaptability Dynamically adjust model parameters: Using incremental least squares methods and sliding window methods, model parameters are updated in real time, allowing the model to adapt to the dynamic characteristics of crack changes. When the crack growth rate or direction changes, the model parameters can be automatically adjusted to maintain the model's predictive performance.

[0110] Adaptability to different crack types and environmental conditions: By establishing a multi-factor association model and a machine learning model, the model parameters can be adjusted according to different crack types (such as opening cracks, sliding cracks) and different environmental conditions (such as different temperature ranges and humidity levels), making the model more versatile and adaptable.

[0111] 3. Improve monitoring efficiency Real-time monitoring and early warning: Combining real-time data updates with predictive models enables real-time crack monitoring and early warning. When the prediction results indicate that the width or length of a crack is about to exceed the safety threshold, a warning signal is issued to prompt maintenance personnel to take measures to prevent further crack expansion and potential safety accidents.

[0112] 4. Optimize maintenance decisions Predicting crack growth trends: Accurately predicting crack expansion trends provides a scientific basis for maintenance decisions. For example, by predicting the width of cracks over time, maintenance personnel can plan maintenance in advance and choose the appropriate maintenance timing and methods, avoiding excessive or untimely maintenance.

[0113] Extending the service life of the structure: By promptly detecting and addressing crack problems, the cracks can be prevented from further expanding, thereby extending the service life of the lock metal structure and reducing maintenance costs.

[0114] In summary, the newly added crack development trend prediction method significantly improves prediction accuracy, enhances model adaptability, improves monitoring efficiency, optimizes maintenance decisions, and improves system reliability through real-time data updates, comprehensive application of multiple models, and consideration of multiple factors. It provides strong technical support for the safe monitoring and maintenance of ship lock metal structures.

[0115] The contents described in the embodiments of this specification are merely an enumeration of the implementation forms of the inventive concept. The scope of protection of the present invention should not be regarded as limited to the specific forms described in the embodiments. The scope of protection of the present invention also extends to equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.

Claims

1. A ship lock metal structure detection system based on high-precision grating strain gauge technology, characterized by: It includes an outer shell and multiple first oil cylinders, multiple valve blocks and multiple second oil cylinders are fixedly arranged in the outer shell, a stepped hole is provided in the valve block, a shaft tube is assembled in the fine hole in the stepped hole, the shaft tube is located outside the fine hole and is fixedly connected to the top pressure frame, a clamping sleeve is threadedly connected to one end of the coarse hole in the stepped hole, a first spring is arranged on the outer sleeve of the shaft tube, the two ends of the first spring respectively rest on the valve block and the top pressure frame, the piston rod of the second oil cylinder is fixedly connected to the top pressure frame, a grating strain gauge is provided in the stepped hole, and the two ends of the grating strain gauge are respectively fixedly connected to the clamping sleeve and the shaft tube by clamping parts; the two ends of the first oil cylinder are respectively connected to a hinge seat, the first oil cylinder is connected to the second oil cylinder through an oil pipe, and when the same hydraulic oil is input to the first oil cylinder and the second oil cylinder, the piston rod stroke of the first oil cylinder is smaller than the piston stroke of the second oil cylinder.

2. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to claim 1 is characterized by: A collecting pipe is arranged in the shell. The collecting pipe has an oil discharge port and a plurality of oil inlets. Each first oil cylinder is connected with each oil inlet through an oil pipe, and the oil discharge port is connected with the second oil cylinder through an oil pipe.

3. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to claim 1 is characterized in that The model of the grating strain gauge is FGB-1000, with a measuring range of 10 mm and an accuracy of 0.01 mm.

4. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to claim 1 is characterized in that :The pistons of the first and second oil cylinders are made of high-strength aluminum alloy.

5. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to any one of claims 1 to 4, characterized in that: The shell includes a shell and a shell cover, the shell and the shell cover are sealed together, a plurality of through-wall straight-through pipe joints are installed on the shell, a seal is provided between the through-wall straight-through pipe joints and the shell, the oil pipes inside and outside the shell are connected to both ends of the through-wall straight-through pipe joints, a plurality of waterproof joints are installed on the shell, and the optical fiber connected to the grating strain gauge is assembled in the waterproof joint.

6. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to any one of claims 1 to 4, characterized in that: The first oil cylinder is a spring return oil cylinder.

7. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to any one of claims 1 to 4, characterized in that: The clamping part includes a mounting hole opened at one end of the ferrule and the shaft tube, and the mounting hole includes a countersunk hole, a tapered hole and a threaded hole. A second spring is provided in the countersunk hole, and a plurality of fan-shaped tiles are provided in the tapered hole. The multiple fan-shaped tiles are matched with the tapered hole in an inclined surface. The threaded hole is threadedly connected with a clamping sleeve, and the clamping sleeve pushes the fan-shaped tiles against the second spring.

8. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to any one of claims 1 to 4, characterized in that: It also includes a plurality of pipe clamps, which include clamping blocks. The bottom surface of the clamping block is provided with a clamping groove for fixing the oil pipe, and magnets are fixedly connected to the clamping block on both sides of the clamping groove.

9. The ship lock metal structure detection system based on high-precision grating strain gauge technology according to any one of claims 1 to 4, characterized in that: It also includes a signal transmitter, a multi-channel data conversion module, a data demodulator and a data analysis and processing module. Each grating strain gauge transmits the collected signal to the multi-channel data conversion module after passing through the signal transmitter. The multi-channel data conversion module transmits the signal to the data demodulator. The data demodulator converts the analog signal into a digital signal and sends it to the data analysis and processing module. After the data analysis and processing module processes the data, the data is displayed on the display.

10. A method for monitoring cracks using the ship lock metal structure detection system based on high-precision grating strain gauge technology as claimed in claim 9, characterized in that: When the cracks on both sides of the first cylinder change, the first cylinder moves along with the cracks, the second cylinder amplifies the movement of the first cylinder, drives the grating strain gauge to expand and contract, causing the wavelength of the grating strain gauge to change, and the grating strain gauge sends the change signal to the signal transmitter. The signal transmitter processes the signal transmitted by the grating strain gauge and transmits it to the multi-channel data conversion module. The multi-channel data conversion module transmits the signal to the data demodulator. The data demodulator converts the analog signal into a digital signal and sends it to the data analysis and processing module. The data analysis and processing module processes and stores the data and displays the crack status on the display.

11. The crack monitoring method based on high-precision grating strain gauge technology according to claim 10, characterized in that The data analysis and processing module processes and stores the data and predicts the crack development trend. The prediction method includes the following steps: S1: Data collection and preprocessing: Collect crack detection data, including crack width y i and length, as well as related environmental conditions and lock operation status; pre-process the collected data, including data cleaning and normalization. The specific formula is: Where x is the original data, x min and x max are the minimum and maximum values of the data column respectively; S2: Establish crack development trend model: Establish a crack development trend model using a linear regression model: y=a+bt Where a is the intercept and b is the slope, which represent the crack growth rate. The parameters a and b are estimated by the least squares method: where t i and y i are the time and crack width of the i-th measurement, and are the average values of time t and crack width y, respectively; S3: Predict future crack width: Use the established model to predict the future time point t future The crack width y future : yes future =a+bt future S4: Update model parameters in real time: When a new data point (t new ,y new ) arrives, calculate the residual: residual=y new -(a+bt new ) Adjust model parameters based on residuals: Δa=learning_rate×residual Δb=learning_rate×residual×t ne w Among them, residual represents the residual of the new data point, tnew is the time of the new data point, and learning_rete is the learning rate, which is used to control the step size of parameter update; Update model parameters: in new =a+Δa b new =b+Δb Δa represents the adjustment amount of the model intercept a, which is used to adjust the intercept of the model to better fit the new data points; Δb represents the adjustment amount of the model slope b, which is used to adjust the slope of the model to fit the new data points.

12. The crack monitoring method based on high-precision grating strain gauge technology according to claim 11, characterized in that In step S4, the sliding window method is used to update the parameters. The method is as follows: a fixed-size data window is maintained, and when a new data point (t new ,y new ) arrives, the oldest data point in the window is moved, and the data is added to the window, and the model parameters a and b are recalculated using the data in the window.

Images