A smart ground-to-air laying machine and a method for laying power optical cables

Abstract

This invention belongs to the technical field of power construction equipment, specifically relating to an integrated intelligent ground-to-air laying machine and method for power optical cables. The integrated intelligent ground-to-air laying machine includes an intelligent optical cable laying machine, a cable threader, and a traction medium unwinding machine. The traction medium unwinding machine is connected to the cable threader via a fixed hook. The cable threader contains the power optical cable to be laid, which is connected to the intelligent optical cable laying machine. The intelligent optical cable laying machine has upper and lower limit rollers at both its beginning and end, through which the power optical cable passes. This invention uses a BP neural network to predict the optimal traction torque and traction speed for laying the power optical cable under current working conditions. It can automatically adjust laying parameters within the optimal range, avoiding over-traction of all-dielectric self-supporting optical cables and ordinary cables, reducing the labor intensity of on-site workers, and improving the inherent safety of power optical cable operation.

Description

Technical Field

[0001] This invention belongs to the field of power construction equipment technology, specifically relating to an integrated machine for intelligent ground and air laying of power optical cables and a method for laying power optical cables. Background Technology

[0002] With the continuous improvement of the goals for building new power systems, the scale and quality requirements for power grid construction are increasing. Traditional power grid construction methods have many shortcomings in terms of safety, labor consumption, and operational efficiency, making it difficult to adapt to the ever-changing demands. Especially during power grid maintenance and construction, traditional manual operations not only consume a large amount of manpower but also have low efficiency and pose significant safety risks. Therefore, power grid construction urgently needs to improve work efficiency and safety through mechanized construction methods to promote high-quality development of power grid construction.

[0003] Fiber optic cable laying is a crucial component of power construction, particularly in the cable traction stage, which often requires significant manual labor. Especially when laying cables within ducts, traditional cable conveyors are difficult to handle, noisy, and unsuitable for underground operations due to their size and other limitations. Therefore, in these working environments, traction media (such as Dyneema rope, steel wire rope, and conduit pullers) often require manual deployment and retraction, resulting in high labor intensity and low efficiency.

[0004] Chinese patent application CN208818873U discloses an electric optical cable traction machine, belonging to the technical field of optical cable laying equipment. It includes a base frame, support frame, drag handle, and housing; the traction device includes a motor, reduction gearbox, drive gear, reverse gear, an upper belt on the drive roller, a driven roller below the drive roller with a lower belt on the driven roller, a power interface, power switch, frequency converter, and high / low speed switch on the side of the housing; a counter device, panel button device, receiving antenna, dedicated button for cable threading, and emergency stop button are located on the top of the housing; spring locking structures are provided on both sides of the drive roller's roller support, and front and rear limiters are also provided on both sides respectively; a guide wheel is located below the front limiter, and a guide connector is also provided on the roller support; ground spikes are also provided on the side of the housing. The above device cannot achieve intelligent control of torque and traction speed, and cannot deploy and retrieve the traction medium, resulting in low automation. Therefore, it is urgent for those skilled in the art to solve the above technical problems. Summary of the Invention

[0005] The technical problem to be solved by the present invention is that the prior art cannot achieve intelligent control of torque and traction speed, and cannot deploy and recover the traction medium, resulting in a low degree of automation.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0007] A smart cable laying machine for power optical cables, characterized in that it includes a smart cable laying machine, a cable threader, and a traction medium unwinding machine. The traction medium unwinding machine is connected to the cable threader via a fixed hook. The cable threader contains the power optical cable to be laid. The power optical cable is connected to the smart cable laying machine. The smart cable laying machine has upper limit wheels and lower limit wheels at both its head and tail ends. The power optical cable passes through the upper limit wheels and the lower limit wheels.

[0008] The integrated intelligent optical cable laying machine is equipped with a control cabinet. The control cabinet controls the operating parameters of the integrated intelligent optical cable laying machine through a built-in PLC and servo controller to ensure the stable laying of power optical cables.

[0009] By adopting the above technical solution, and through the coordinated operation of the traction medium unwinding machine and the integrated intelligent optical cable laying machine, the efficiency of optical cable laying is greatly improved. The traction medium unwinding machine can automatically unwind the traction medium and work synchronously with the integrated intelligent optical cable laying machine, avoiding the low efficiency and high labor intensity of traditional manual unwinding. The equipment has a high degree of automation, especially in the automatic unwinding and rewinding function of the traction medium, which significantly reduces manual intervention and lowers the labor intensity of construction workers, especially important in laying operations that require long-term, high-intensity operation. In addition, the integrated intelligent optical cable laying machine, through the design of upper and lower limit wheels, ensures that the optical cable maintains a stable trajectory during the laying process, avoiding deviation or damage caused by excessive or insufficient traction force. The equipment also effectively protects the optical cable, preventing it from being subjected to excessive tension or impact during the laying process, thus avoiding damage. Combined with the cable puller and the integrated intelligent optical cable laying machine, the equipment can adapt to various working environments, especially in pipeline laying and underground operations, overcoming the problems of difficult handling and high noise of traditional equipment.

[0010] Furthermore, the integrated intelligent optical cable laying machine also includes an adjusting handwheel, a transmission belt, a meter counting wheel, a transmission wheel, a fixed bracket, a drag roller, a tensioning wheel, a power supply, a control cabinet, a guide trolley, and a snap-on slide rail;

[0011] The upper limit wheel and the lower limit wheel are fixed to the fixed bracket by bolts;

[0012] The adjusting handwheel is connected to the fixed bracket and the idler roller via a connecting shaft, and the conveyor belt is connected to the transmission wheel and the meter counting wheel / detection wheel via a pulley and the tension wheel.

[0013] The measuring wheel is fixed to the transmission belt and the fixed bracket by bearings, and the transmission wheel is connected to the fixed bracket by bearings.

[0014] The idler roller is connected to the fixed bracket via a bearing, and the tensioning wheel is mounted on the fixed bracket;

[0015] The lower limit wheel is fixed to the fixed bracket by bolts, the power supply is connected to the control cabinet by a cable, and the control cabinet is fixed together with the fixed bracket and the power supply by bolts.

[0016] The guide trolley is fixed to the bracket by a snap-on slide rail, and the snap-on slide rail is fixed to the fixed bracket and the guide trolley by bolts.

[0017] By adopting the above technical solutions, the intelligent optical cable laying machine ensures stable transmission of optical cables during the laying process through the cooperation of the drive belt, drive wheel, and meter-counting detection wheel. The meter-counting detection wheel can accurately measure the laying length of the optical cable and record it accurately during forward and reverse rotation, avoiding the inaccuracies of traditional manual measurement, thereby improving work efficiency and data accuracy. The upper and lower limit wheels are fixed to the fixed bracket with bolts to ensure that the optical cable maintains a stable trajectory during the laying process, avoiding the risk of optical cable deviation, knotting, or damage. The design of the idler rollers, drive wheel, and tensioning wheel further enhances the stability of the equipment, ensuring that the optical cable is not prone to jamming or instability during transmission. The adjustment handwheel is connected to the fixed bracket and idler rollers, providing a simple adjustment method that can easily adjust the traction and tension according to different optical cable materials and laying environments, ensuring stable traction and smooth optical cable transmission throughout the laying process. In addition, the bearing connection design of the tensioning wheel and meter-counting detection wheel makes equipment maintenance and upkeep more convenient and faster.

[0018] Furthermore, the traction medium unwinding machine includes an operating lever, a telescopic joint, a machine compartment, a drive wheel, a frame, an adjusting handwheel, a trolley wheel, an external power supply, a fixing hook, and a control panel;

[0019] The operating lever is fixed to the frame with bolts, the telescopic joint is installed in the center of the operating lever, the cabin is fixed to the frame with bolts, the drive wheel is connected to the frame through bearings, and the frame is connected to the operating lever, the cabin, and the fixing hook with bolts.

[0020] The adjusting handwheel is connected to the frame and the fixing hook via a connecting shaft, the small wheel is installed at the bottom of the frame, and the control panel is fixed to the top of the cabin.

[0021] By adopting the above technical solution, components such as the operating lever, telescopic joint, engine compartment, drive wheel, and frame are connected by bolts and bearings, ensuring stable connections between components and enhancing the overall stability of the equipment. In particular, the drive wheel's connection to the frame via bearings effectively guarantees the smooth operation of the drive system, avoiding vibration or instability during operation and thus ensuring the long-term reliability of the equipment. The telescopic joint design allows the length of the operating lever to be adjusted according to different operational needs, enhancing the equipment's adaptability to various operating environments. Especially in confined spaces or when adjusting the working angle, the telescopic nature of the operating lever facilitates flexible adjustments by the operator, simplifying on-site operation. The control panel is installed on the top of the engine compartment, allowing operators to monitor and adjust the equipment's operating status at any time, simplifying the operation process. The connection between the adjusting handwheel and the frame and fixed hook allows for convenient and quick adjustment of the tension and position of the traction medium, ensuring precise control of the traction force.

[0022] Furthermore, the upper limit wheel and the lower limit wheel are provided with U-shaped grooves;

[0023] The adjusting handwheel is equipped with a linkage crossbar, which is connected to the fixing hook, and the fixing hook is located at the bottom of the traction medium unwinding machine.

[0024] By adopting the above technical solution, the adjusting handwheel is equipped with a linkage crossbar, which is connected to the fixed hook. The traction force and fiber optic cable tension can be easily adjusted by adjusting the handwheel. This design allows operators to precisely control the traction force and tension of the fiber optic cable, thereby avoiding over-traction or under-traction and ensuring that the fiber optic cable is not damaged during laying. The linkage crossbar design enables the adjustment handwheel and the fixed hook to work together, facilitating rapid response to the needs of different working environments. Through simple operation, the tension of the fiber optic cable can be quickly adjusted, making the traction process more efficient and precise. The fixed hook is located at the bottom of the traction medium unwinding machine, and its connection with the linkage crossbar automates the unwinding and rewinding process of the traction medium, reducing the need for manual intervention and simplifying the operation process. This significantly reduces labor intensity and operational complexity for laying work that requires long-term operation.

[0025] Furthermore, the measuring wheel is equipped with a tension spring, the measuring wheel is mounted on the side wall of the transmission belt, the fixing bracket is made of aluminum alloy, and the tension wheel is mounted on the transmission belt.

[0026] By adopting the above technical solution, the measuring wheel is equipped with a tension spring, which maintains close contact between the measuring wheel and the optical cable, avoiding measurement errors caused by loosening or slipping of the optical cable. The tension spring design ensures that the measuring wheel maintains stable measurement accuracy under different loads, thereby ensuring accurate length recording during the optical cable laying process. The measuring wheel is mounted on the side wall of the drive belt, ensuring precise matching between the measuring wheel and the drive belt and avoiding friction or misalignment problems caused by improper installation position. This design ensures that the measuring wheel can stably measure the laying length of the optical cable, while making the transmission of the optical cable smoother during the laying process and reducing resistance. The fixing bracket is made of aluminum alloy, which has strong load-bearing capacity, corrosion resistance, and lightweight. The aluminum alloy material not only improves the stability and durability of the equipment, but also maintains good working performance during long-term use, reduces wear and tear in harsh environments, and extends the service life of the equipment. The tension wheel, mounted on the drive belt, can effectively adjust the tension of the drive belt and ensure smooth transmission. By precisely controlling the tension of the drive belt, it is possible to ensure that the optical cable maintains appropriate tension during the laying process, preventing damage or operational instability caused by the optical cable being too tight or too loose.

[0027] This invention also discloses a method for laying power optical cables, which utilizes an integrated intelligent ground and air laying machine for power optical cables to achieve the laying of power optical cables, specifically including the following steps:

[0028] S1: First, install the traction medium unwinding machine, and pass the fixing hook of the traction medium unwinding machine through the cable threader. The cable to be laid through the cable threader passes through the optical cable intelligent laying integrated machine and is clamped.

[0029] S2: Initialize the traction medium unwinding machine through the control panel to ensure that the unwinding machine is in standby mode and ready to start the unwinding operation;

[0030] S3: Initialize the intelligent optical cable laying machine through the control cabinet and input the laying parameters of the power optical cable. The intelligent optical cable laying machine uses a BP neural network trained in the PLC system to predict the traction torque range and traction speed range.

[0031] S4: Start the traction medium unwinding machine and the integrated intelligent optical cable laying machine. The power optical cable is pulled by the intelligent laying machine, and the traction medium unwinding machine simultaneously unwinds the traction medium.

[0032] S5: The laying of the power optical cable begins, and the traction medium unwinding machine simultaneously recovers the traction medium;

[0033] S6: During the laying process, the PLC system built into the control cabinet verifies the laying parameter information output in real time by the integrated intelligent optical cable laying machine to ensure the accuracy of the operation.

[0034] By adopting the above-mentioned technical solutions, the extensive application of automation systems during optical cable laying, especially the automatic unwinding and retrieval of the traction medium, significantly reduces manual intervention and lowers the labor intensity of construction workers. This advantage is particularly prominent in long-term, high-intensity laying operations. Through the neural network prediction of the intelligent optical cable laying machine, the optimal traction torque and traction speed can be accurately calculated and controlled, avoiding the errors and instabilities of traditional manual control. This ensures precise adjustment of the traction force, preventing over-traction or damage to the optical cable, thereby improving operational accuracy and cable safety. Throughout the laying process, the equipment design ensures that the optical cable is pulled with appropriate tension, preventing damage due to excessive traction. Simultaneously, the synchronous retrieval function of the traction medium unwinding machine reduces excess tension in the traction medium, further protecting the optical cable. Furthermore, the real-time monitoring and calibration system ensures the accuracy of laying parameters (such as traction force and laying speed), enabling timely detection of problems and preventing cable damage or operational difficulties. The equipment design allows it to adapt to different operating environments, especially in complex environments such as underground and pipeline locations, ensuring efficient and stable operation and overcoming the problems of difficult handling and high noise levels associated with traditional equipment. The intelligent control system and real-time data monitoring reduced experience-based errors in manual operations, ensuring the safe laying of optical cables. In particular, it improved construction safety by avoiding over-traction and other potential hazards. The introduction of automation and intelligence enabled efficient execution of each work step, reducing unnecessary adjustments and accelerating the overall progress of the optical cable laying project.

[0035] Further, in step S1, the adjusting handwheel of the traction medium unwinding machine is rotated to engage the fixing hook with the threader;

[0036] In step S3, by inputting the laying length, material, diameter, ground and air markings, and traction medium quantification parameters of the power optical cable, the optimal traction torque and traction speed of the optical cable conveyor are predicted by the BP neural network.

[0037] By adopting the above technical solution

[0038] Furthermore, the BP neural network employs forward propagation and backward propagation algorithms, and consists of an input layer, a hidden layer, and an output layer. The input is a column vector provided by data samples, which is multiplied by each connection matrix to obtain the neurons of each node. Substituting the node values ​​of each neuron into the transfer function yields the desired output result.

[0039] In the input layer, x1, x2, ... xi, xn are used as inputs, and in the output layer, y1, y2, ym are used as outputs. The input vector set of the BP neural network is {(x, y)}.

[0040] In the hidden layers, let n be the number of input layers, α be the adjustment constant, and m be the number of output layers. Then the number of hidden layers is:

[0041]

[0042] Forward propagation is the process of a signal moving from the input layer to the hidden layer and then to the output layer. Let the connection parameters of the neuron be ωji, the threshold be bi, f(s) be the activation function, xi be the neuron input, and yj be the neuron output, then the expression is as follows:

[0043] sj=x1*ωi1+bj1+x2*ωj2+bj2+…x1*ωji+bji+…+xn*ωjn+bjn;

[0044] yj = f(sj);

[0045] Backpropagation is the process by which error travels from the output layer to the hidden layers and then back to the input layer. Gradient descent is used to adjust the connection weights and bias terms in the network to reduce the error between the network output and the desired output. The error function is the mean squared error, let d k For the desired output, o k For the actual output, the mean squared error is:

[0046]

[0047] The weight gradient error is calculated using the chain rule, and the weight adjustment formula for the connection parameters is:

[0048] Where η is the learning rate;

[0049] The output layer uses the sigmoid function as the activation function, with the output ranging from [0, 1]. The expression for the sigmoid function is:

[0050] By adopting the above technical solution, and by inputting parameters such as the length, material, diameter, and ground / air markings of the optical cable, combined with BP neural network prediction, the optimal traction torque and traction speed can be calculated in real time. This precise control not only improves the safety of the laying operation but also avoids the errors and instabilities of traditional manual control. Through forward and backward propagation algorithms, the BP neural network can automatically adjust the connection weights and bias terms, enabling the network to gradually reduce errors and optimize control parameters. This makes the fiber optic cable laying process more flexible and adaptive, capable of adapting to different types of fiber optic cables and different working environments. The automated learning and adjustment process of the network reduces subjective judgment bias in manual operation, thereby reducing the occurrence of human error and improving the accuracy and safety of fiber optic cable laying. The BP neural network uses the gradient descent method to optimize parameters, effectively adjusting the connection weights in the network, thereby enhancing the stability of the system and avoiding work interruptions or fiber optic cable damage caused by inaccurate parameters or improper adjustments. Through automated calculation and adjustment, manual intervention is reduced and work efficiency is improved. Especially in long-term and high-intensity laying operations, this technology can significantly save time and improve work progress. Based on the prediction and adjustment function of the neural network, this solution can adapt to a variety of complex working environments, ensuring stable operation of the equipment and overcoming problems such as difficult handling and high noise that may exist in traditional equipment, further improving the reliability and efficiency of fiber optic cable laying.

[0051] Furthermore, the training process of the BP neural network includes the following steps:

[0052] S31: Determine the number of hidden layer nodes and adjust the weight matrix of the BP model;

[0053] S32: The number of hidden layer nodes is determined, and the number of hidden layer neurons is used as an input parameter to test the same sample set data;

[0054] The optimal number of neurons in the hidden layer can be determined by the following steps:

[0055] S311: Testing from a small number of nodes;

[0056] S312: Conduct network testing and training;

[0057] S313: Test with increased node count;

[0058] S314: Determine the number of nodes;

[0059] During training, the network's generalization ability is increased, interference factors are reduced, and samples with the same number of iterations are selected for testing with different numbers of hidden nodes. The resulting error and maximum number of iterations are used as input parameters for further testing to obtain the model with the optimal number of hidden nodes.

[0060] By adopting the above technical solution and optimizing the number of hidden layer nodes, overfitting or underfitting can be avoided, ensuring the model has a good fit to the training data and improving generalization ability. This allows the model to maintain high prediction accuracy under different types of optical cables, materials, and ground / air markings. This optimization process reduces the impact of external interference, improves network stability, and ensures consistent performance in different environments. By dynamically adjusting the number of hidden layer nodes and conducting multiple tests, the model can adapt to the needs of different operating environments, thereby achieving optimal traction torque and traction speed predictions. Gradually increasing the number of nodes and conducting tests also efficiently trains the network, shortens training time, and improves work efficiency and safety. Ultimately, this training method improves the stability and reliability of the network, ensuring consistent and accurate prediction results under complex operating conditions, thereby improving the overall quality and safety of optical cable laying.

[0061] Furthermore, the initialization of the traction medium unwinding machine in step S2 includes the following steps:

[0062] S21: Initialize the network, assign random values ​​in the interval [-1, 1] to the connection parameter weights ω and the threshold b, and calculate the number of hidden layers p based on the number of input layers, output layers and adjustment constant a;

[0063] S22: Training model assignment, providing the learning pattern pair [Xk Yk] composed of power optical cable laying parameters to the neural network;

[0064] S23: Forward propagation signal, sequentially calculate the net input and output of each unit in the hidden layer and output layer.

[0065] S24: Backpropagation error, calculate the expected output dj and mean square error E, and calculate the generalized error of each unit in the output layer and hidden layer in turn;

[0066] S25: Adjust the connection weights between the hidden layer and the output layer, and between the input layer and the hidden layer, as well as the threshold values ​​of each unit in the output layer;

[0067] S26: Randomly provide the next optional learning data pair to the BP neural network, and return to step S23 until all learning data pairs have been trained;

[0068] S27: Determine whether the global error of the network model can reach the required accuracy, i.e. whether it is valid. If it is valid, jump to step S28. If it is not valid, return to step S23 if the number of learning cycles is sufficient, otherwise continue to step S28.

[0069] S28: The BP neural network is input with the power optical cable laying length, material, diameter, ground and air markings, and traction medium quantification parameters, and completes the normalization process;

[0070] S29: The BP neural network outputs the corresponding optimal traction torque and traction speed, and restores the real data.

[0071] The present invention has the following beneficial effects:

[0072] 1. This invention achieves fully automated operation by combining the intelligent optical cable laying machine and the traction medium unwinding machine, and by using a BP neural network to intelligently predict and adjust the traction torque and traction speed. Through real-time calculation and precise adjustment, it avoids errors and instability in manual operation, significantly improves the efficiency and safety of laying operations, and greatly reduces the need for manual intervention, especially in high-intensity, long-term operations.

[0073] 2. This invention, by employing a meter-counting wheel and a transmission belt, ensures stable transmission of the optical cable during laying and measures the laying length in real time, avoiding errors associated with traditional manual measurement. Combined with the learning and optimization process of a BP neural network, it can automatically adjust the optimal traction torque and traction speed according to the parameters of the optical cable, ensuring the accuracy of optical cable laying, reducing human error, and improving work efficiency.

[0074] 3. The intelligent ground and air laying machine of the present invention is designed to be adaptable to different working environments. Especially in complex environments, it overcomes the difficulties in handling and noise problems of traditional equipment through efficient and stable operation. Through intelligent control system and real-time monitoring, it avoids over-traction and damage to optical cables during the laying process, effectively improves the safety of the construction process, and ensures the stable operation of the equipment and the inherent safety of the optical cable. Attached Figure Description

[0075] Figure 1 This is a schematic diagram of the structure of the intelligent optical cable laying integrated machine of the present invention;

[0076] Figure 2 This is a schematic diagram of the structure of the traction medium unwinding machine of the present invention;

[0077] Figure 3 This is a diagram of the BP neural network structure of the present invention;

[0078] Figure 4 This is a graph of the sigmoid function of the present invention;

[0079] Figure 5 This is the control flowchart of the present invention;

[0080] Figure 6 This is a flowchart of the speed control process of the present invention;

[0081] Figure 7 This is a flowchart of the method of the present invention;

[0082] Figure 8 This is a schematic diagram of the overall connection of the present invention.

[0083] Among them, 1-Intelligent optical cable laying integrated machine; 11-Adjusting handwheel one; 12-Transmission belt; 13-Meter counting wheel; 131-Tension spring; 14-Transmission wheel; 15-Fixed bracket; 16-Drag roller; 17-Tension wheel; 18-Power supply; 19-Control cabinet; 20-Guide trolley; 21-Snap-on slide rail; 2-Threading device; 3-Traction medium unwinding machine; 31-Operating lever; 32-Expansion joint; 33-Cabinet; 34-Drive wheel; 35-Frame; 36-Adjusting handwheel two; 361-Linkage crossbar; 37-Small wheel; 38-External power supply; 39-Fixed hook; 40-Control panel; 4-Power optical cable; 5-Upper limit wheel; 6-Lower limit wheel. Detailed Implementation

[0084] The present invention will now be described in further detail with reference to the accompanying drawings and specific preferred embodiments.

[0085] In the description of this invention, it should be understood that the terms "left side," "right side," "upper part," "lower part," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. "First," "second," etc., do not indicate the importance of the components, and therefore should not be construed as a limitation of this invention. The specific dimensions used in this embodiment are only for illustrating the technical solution and do not limit the scope of protection of this invention.

[0086] like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8As shown, this invention discloses an integrated intelligent ground and air laying machine for power optical cables, including an intelligent optical cable laying machine 1, a cable threader 2, and a traction medium unwinding machine 3. The intelligent optical cable laying machine 1 consists of an upper limit wheel 5 and a lower limit wheel 6, which are bolted to a fixed bracket 15 to ensure that the power optical cable 4 passes stably through the upper limit wheel 5 and lower limit wheel 6 during laying, avoiding deviation. The upper limit wheel 5 and lower limit wheel 6 are provided with U-shaped grooves to more securely guide the power optical cable 4. The intelligent optical cable laying machine 1 also includes an adjusting handwheel 11, a transmission belt 12, a meter counting wheel 13, a transmission wheel 14, a fixed bracket 15, a drag roller 16, a tensioning wheel 17, a power supply 18, a control cabinet 19, a guide trolley 20, and a snap-on slide rail 21. The adjusting handwheel 11 is connected to the fixed bracket 15 and the drag roller 16 via a connecting shaft and is used to adjust the tension and traction of the power optical cable 4. The transmission belt 12 is connected to the transmission wheel 14 and the meter-counting detection wheel 13 via the tensioning pulley 17, ensuring smooth transmission of the power optical cable 4. The meter-counting detection wheel 13 is used to accurately measure the length of the laid power optical cable 4 and provide real-time data to the control system to ensure the accuracy of the laying of the power optical cable 4. The traction medium unwinding machine 3 includes an operating lever 31, a telescopic joint 32, a machine compartment 33, a drive wheel 34, a frame 35, an adjusting handwheel 36, a small wheel 37, an external power supply 38, a fixing hook 39, and a control panel 40. The operating lever 31 is fixed to the frame with bolts. The telescopic joint 32 allows the equipment to adjust the height and angle of the traction medium unwinding machine 3 according to the operation requirements. The adjusting handwheel 36 is connected to the frame 35 and the fixing hook 39 via a connecting shaft, facilitating the adjustment of the tension of the traction medium. The fixing hook 39 is located at the bottom of the traction medium unwinding machine 3, facilitating the connection of the cable threader 2. The small wheel 37 is installed at the bottom of the frame 35, improving the mobility of the equipment and facilitating movement in the working environment. The traction medium unwinding machine 3 is connected to the cable threader 2 via a fixed hook 39. The cable threader 2 contains the power optical cable 4 to be laid. The power optical cable 4 is transmitted to the intelligent optical cable laying machine 1 via the upper limit wheel 5 and the lower limit wheel 6, where it is pulled and controlled. The machine is connected to a power supply and control cabinet 19 to monitor and adjust the equipment's operating status in real time, ensuring efficient and safe operation.

[0087] Working Principle: Before starting the equipment, the traction medium unwinding machine 3 is first adjusted to a suitable height using the operating lever 31, and then connected to the bottom of the cable threader 2 using the fixing hook 39 to ensure the steady passage of the power optical cable 4. The intelligent optical cable laying machine 1 adjusts the traction force and tension of the power optical cable 4 by adjusting the handwheel 11 to ensure smooth transmission of the power optical cable 4. After the equipment starts, the power optical cable 4 is pulled by the intelligent optical cable laying machine 1, while the traction medium unwinding machine 3 unwinds the traction medium synchronously. The transmission belt 12 and the meter counting wheel 13 ensure the accurate transmission of the power optical cable 4 and measure its length. The meter counting wheel 13 provides real-time data to the control system for accurately calculating the laying progress of the power optical cable 4. During the laying process, the monitoring system verifies the output data of the intelligent optical cable laying machine 1 in real time to ensure that the traction force and speed are adjusted within the optimal range, thereby avoiding over-traction or damage to the optical cable.

[0088] The present invention further discloses a method for laying power optical cables, which utilizes an integrated intelligent ground and air laying machine for laying power optical cables to achieve the laying of power optical cables 4, specifically including the following steps:

[0089] S1: First, install the traction medium unwinding machine 3, and pass the fixing hook 39 of the traction medium unwinding machine 3 through the cable threader 2. The cable to be laid by the cable threader 2 passes through the optical cable intelligent laying integrated machine 1 and is clamped.

[0090] S2: Initialize the traction medium unwinding machine 3 to ensure that the unwinding machine is in standby mode and ready to start the unwinding operation;

[0091] S3: Initialize the intelligent optical cable laying machine 1 and input the laying parameters of the power optical cable 4. The intelligent optical cable laying machine 1 uses a BP neural network to predict the optimal traction torque and traction speed.

[0092] S4: Start the traction medium unwinding machine 3 and the optical cable intelligent laying machine 1. The power optical cable 4 is pulled by the intelligent laying machine, and the traction medium unwinding machine 3 simultaneously unwinds the traction medium.

[0093] S5: The laying of the power optical cable 4 begins, and the traction medium unwinding machine 3 simultaneously recovers the traction medium;

[0094] S6: During the laying process, the laying parameter information output in real time by the intelligent optical cable laying machine 1 is checked to ensure the accuracy of the operation.

[0095] In step S1, rotate the adjusting handwheel 36 of the traction medium unwinding machine 3 to engage the fixing hook 39 with the threader 2;

[0096] In step S3, by inputting the laying length, material, diameter, ground and air markings, and traction medium quantification parameters of the power optical cable 4, the optimal traction torque and traction speed of the optical cable conveyor are predicted by the BP neural network.

[0097] The BP neural network uses forward and backward propagation algorithms and consists of an input layer, a hidden layer, and an output layer. The input is a column vector provided by the data samples. Multiplying the vector by each connection matrix yields the neurons at each node. Substituting the node values ​​of each neuron into the transfer function yields the desired output result.

[0098] In the input layer, x1, x2, ... xi, xn are used as inputs, and in the output layer, y1, y2, ym are used as outputs. The input vector set of the BP neural network is {(x, y)}.

[0099] In the hidden layers, let n be the number of input layers, α be the adjustment constant, and m be the number of output layers. Then the number of hidden layers is:

[0100]

[0101] Forward propagation is the process of a signal moving from the input layer to the hidden layer and then to the output layer. Let the connection parameters of the neuron be ωji, the threshold be bi, f(s) be the activation function, xi be the neuron input, and yj be the neuron output, then the expression is as follows:

[0102] sj=x1*ωi1+bj1+x2*ωj2+bj2+…x1*ωji+bji+…+xn*ωjn+bjn;

[0103] yj = f(sj);

[0104] Backpropagation is the process by which error travels from the output layer to the hidden layers and then back to the input layer. Gradient descent is used to adjust the connection weights and bias terms in the network to reduce the error between the network output and the desired output. The error function is the mean squared error, let d k For the desired output, o k For the actual output, the mean squared error is:

[0105]

[0106] The weight gradient error is calculated using the chain rule, and the weight adjustment formula for the connection parameters is:

[0107] Where η is the learning rate;

[0108] The output layer uses the sigmoid function as the activation function, with the output ranging from [0, 1]. The expression for the sigmoid function is:

[0109]

[0110] During training, the network is first initialized by randomly initializing connection weights and bias terms within the range [-1, 1] to ensure the network's initial state is not biased towards any particular direction. Next, the network is trained using fiber optic cable laying data provided in the training set, with various parameters of the fiber optic cable (such as length, material, and traction medium quantization) input for calculation. The output is calculated via forward propagation and compared with the expected output. Then, the error is calculated using the backpropagation algorithm, and gradient descent optimization is performed to update the weights and bias terms. During training, the process ends when the set number of iterations or error accuracy requirements are reached, and the network model reaches its optimal state. In actual fiber optic cable laying operations, the trained BP neural network can predict the optimal traction torque and traction speed in real time based on the specific parameters of the fiber optic cable, thereby automatically adjusting the working state of the fiber optic cable laying machine to ensure smooth cable laying. This intelligent control greatly improves the accuracy and safety of fiber optic cable laying, avoiding problems such as cable damage or low work efficiency caused by human error. Specifically, it includes the following steps:

[0111] S21: Initialize the network, assign random values ​​in the interval [-1, 1] to the connection parameter weights ω and the threshold b, and calculate the number of hidden layers p based on the number of input layers, output layers and adjustment constant a;

[0112] S22: Training model assignment, providing the learning pattern pair [Xk, Yk] composed of the power optical cable 4 laying parameters to the neural network;

[0113] S23: Forward propagation signal, sequentially calculate the net input and output of each unit in the hidden layer and output layer.

[0114] S24: Backpropagation error, calculate the expected output dj and mean square error E, and calculate the generalized error of each unit in the output layer and hidden layer in turn;

[0115] S25: Adjust the connection weights between the hidden layer and the output layer, and between the input layer and the hidden layer, as well as the threshold values ​​of each unit in the output layer;

[0116] S26: Randomly provide the next optional learning data pair to the BP neural network, and return to step S23 until all learning data pairs have been trained;

[0117] S27: Determine whether the global error of the network model can reach the required accuracy, i.e. whether it is valid. If it is valid, jump to step S28. If it is not valid, return to step S23 if the number of learning cycles is sufficient, otherwise continue to step S28.

[0118] S28: The BP neural network inputs the laying length, material, diameter, ground and air markings, and traction medium quantification parameters of the power optical cable 4, and completes the normalization process;

[0119] S29: The BP neural network outputs the corresponding optimal traction torque and traction speed, and restores the real data.

[0120] In another embodiment, step S21: Initialize the network. Before training begins, the network's connection weights and bias terms (thresholds) are randomly initialized. During initialization, each connection weight and bias term is assigned a random value within the interval [-1, 1]. Then, the number of hidden layers is calculated based on the input layer (the number of feature values ​​of the optical cable parameters), the output layer (the number of predicted values ​​of traction torque and traction speed), and an adjustment constant (usually related to network complexity). The number of hidden layer nodes is determined experimentally or through optimization methods, typically adjusted based on the complexity of the input and output data to ensure the model's learning and generalization abilities.

[0121] Step S22: Training Model Assignment. After initialization, the optical cable laying parameters (such as length, material, traction medium quantification parameters, etc.) in the training set are organized into learning pattern pairs. Each pair of learning data is input into the neural network as training data for subsequent forward propagation and error calculation.

[0122] Step S23: Forward Propagation. In the network, the input layer data is weighted and then passed to the hidden layer through an activation function (such as sigmoid or ReLU). The neurons in the hidden layer further process the signal and pass the result to the output layer. In each neuron, a weighted summation operation is first performed, and then an output value is generated through the activation function. The forward propagation process passes the signal layer by layer from the input layer to the output layer, ultimately obtaining the network's prediction results (traction torque and traction speed).

[0123] Step S24: Backpropagation of error. Backpropagation is the core of the neural network optimization process. It reduces error by calculating the errors in the output and hidden layers and adjusting the network weights. During backpropagation, the error between the expected output (target value) and the actual output is first calculated, typically using the mean squared error (MSE) as the error function. The error formula is:

[0124]

[0125] Where, d i For the desired output, y i The output is denoted by , where m is the number of output nodes. Then, the output of each neuron is corrected using the error gradient, and the error signal is propagated back from the output layer to the hidden layer, and finally to the input layer, using a chain rule.

[0126] Step S25: Adjust connection weights and thresholds. Based on the error signal calculated by the backpropagation algorithm, update the connection weights and bias terms (thresholds) between each layer. The weight update rule typically uses gradient descent, and the formula is:

[0127]

[0128] Here, η is the learning rate, which controls the step size of the updates. Through multiple iterations, the network gradually reduces the output error and optimizes the connection weights and bias terms;

[0129] Step S26: Continue training. The network will randomly provide the next pair of learning data. Continue executing steps S23 and S24, continuously updating the weights and biases through multiple forward and backward propagation iterations until all learning data has been trained. After each iteration, the network will gradually optimize and reduce the error.

[0130] Step S27: Determine the stopping condition. During each iteration, the system calculates the global error and compares it with the set accuracy standard. If the global error meets the set accuracy requirement or the maximum number of learning iterations is reached, training stops; if the error still does not meet the requirement, the iteration continues, returning to step S23 for further training.

[0131] Step S28: Data normalization processing. After the network training is completed, the optical cable laying parameters (such as laying length, material, diameter, ground and air markings, and traction medium quantification parameters) input into the BP neural network will first undergo normalization processing. This processing ensures that data with different dimensions and ranges can be effectively processed by the neural network, avoiding the impact of excessively large or small data ranges on the training effect.

[0132] Step S29: Optimal traction torque and traction speed output. The trained and optimized BP neural network outputs the optimal traction torque and traction speed based on the input optical cable laying parameters. These output values ​​will be converted into actual control parameters for use by the intelligent optical cable laying machine 1 to adjust its working state and ensure the smooth progress of the optical cable laying process.

[0133] When controlling the speed of the integrated intelligent optical cable laying machine 1, the master switch of the control cabinet 19 first sends commands such as start, stop, jog, accelerate, and decelerate to the PLC. After receiving the signal, the PLC, as the slave device, calculates the traction and laying frequency (or speed) based on the control commands of the master switch and the calculated corresponding speed and torque. Then, it sends the corresponding control commands to the traction servo drive controller through the RS485 serial bus.

[0134] like Figure 6As shown, the AC servo motor of the integrated intelligent optical cable laying machine 1 adopts closed-loop control. The speed control block diagram is as follows: the traction servo driver receives the control signal from the PLC to drive the servo motor output. The traction force is limited to the optimal traction torque value under the current working condition and is adjusted by the torque limit of the traction motor.

[0135] The AC servo motor control principle of the traction medium unwinding machine 3 is as follows: the unwinding drive adopts closed-loop control. In the working mode of unwinding the traction medium or power optical cable 4, the servo motor adopts constant torque control, and the speed adjustment range is 0-100.2%. In the working mode of rewinding the traction medium, the servo motor adopts constant speed control, and the torque adjustment range is 0-100.2%.

[0136] In actual optical cable laying operations, a trained BP neural network can predict the optimal traction torque and traction speed in real time based on the characteristics of different optical cables and the construction environment, thereby automatically adjusting the working parameters of the optical cable laying machine. Through intelligent control, the equipment can adjust its operating status according to real-time feedback, effectively avoiding problems such as over-traction or optical cable damage, and ensuring the accuracy, efficiency, and safety of optical cable laying operations.

[0137] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A method for laying power optical cables, wherein power optical cables (4) are laid by an intelligent optical cable laying machine (1), wherein the upper limit wheel (5) and lower limit wheel (6) of the intelligent optical cable laying machine (1) are fixed to a fixed bracket (15) by bolts, the adjusting handwheel (11) is connected to the fixed bracket (15) and the roller (16) by a connecting shaft, the transmission belt (12) is connected to the transmission wheel (14) and the meter counting wheel (13) by a tensioning wheel (17), the meter counting wheel (13) and the transmission wheel (14) are fixed to the fixed bracket (15) by bearings, the control cabinet (19) and the power supply (18) are fixed to the fixed bracket (15) by bolts, and the guide trolley (20) is installed on the bracket by a snap-on slide rail (21); the traction medium The operating lever (31) of the medium unwinding machine (3) is fixed to the frame (35) by bolts. The telescopic joint (32) is placed in the center of the operating lever (31). The machine compartment (33) is connected to the frame (35) by bolts. The drive wheel (34) is connected to the frame (35) by bearings. The adjusting handwheel (36) is connected to the frame (35) and the fixed hook (39) by the connecting shaft. The small wheel (37) is installed at the bottom of the frame (35). The control panel (40) is fixed to the top of the machine compartment (33). The cable threader (2) is equipped with the power optical cable (4) to be laid. It is connected by the fixed hook (39) at the bottom of the medium unwinding machine (3). The power optical cable (4) passes through the upper limit wheel (5) and the lower limit wheel (6) of the optical cable intelligent laying integrated machine (1). The characteristic is that: The specific steps involved in laying the power optical cable (4) are as follows: S1: First, install the traction medium unwinding machine (3), and pass the fixing hook (39) of the traction medium unwinding machine (3) through the cable threader (2). The cable to be laid by the cable threader (2) passes through the optical cable intelligent laying machine (1) and is clamped. S2: Initialize the traction medium unwinding machine (3) through the control panel (40) to ensure that the unwinding machine is in standby mode and ready to start the unwinding operation; S3: Initialize the optical cable intelligent laying machine (1) through the control cabinet (19) and input the laying parameters of the power optical cable (4). The optical cable intelligent laying machine (1) uses the BP neural network trained in the PLC system to predict the traction torque range and traction speed range. S4: Start the traction medium unwinding machine (3) and the optical cable intelligent laying machine (1). The power optical cable (4) is pulled by the intelligent laying machine, and the traction medium unwinding machine (3) simultaneously unwinds the traction medium. S5: The power optical cable (4) begins to be laid, and the traction medium unwinding machine (3) simultaneously recovers the traction medium; S6: During the laying process, the PLC system built into the control cabinet (19) calibrates the laying parameter information output in real time by the optical cable intelligent laying machine (1) to ensure the accuracy of the operation process; In step S1, rotate the adjusting handwheel 2 (36) of the traction medium unwinding machine (3) to engage the fixing hook (39) with the threader (2); In step S3, by inputting the laying length, material, diameter, ground and air markings, and traction medium quantification parameters of the power optical cable (4), the BP neural network predicts the traction torque range and traction speed range of the optical cable intelligent laying integrated machine (1).

2. The method for laying power optical cables according to claim 1, characterized in that: The BP neural network uses forward and backward propagation algorithms and consists of an input layer, a hidden layer, and an output layer. The input is a column vector provided by the data samples. Multiplying the vector by each connection matrix yields the neurons at each node. Substituting the node values ​​of each neuron into the transfer function yields the desired output result. In the input layer, x1, x2, ... xi, xn are used as input values, and in the output layer, y1, y2, ym are used as output values. The input vector set of the BP neural network is {(x, y)}. In the hidden layers, let n be the number of input layers, α be the adjustment constant, and m be the number of output layers. Then the number of hidden layers is: Forward propagation is the process of a signal moving from the input layer to the hidden layer and then to the output layer. Let the connection parameters of the neuron be ωji, the threshold be bi, f(s) be the activation function, xi be the neuron input, yj be the neuron output, and sj be the input of the hidden layer neuron. Then the expression is as follows: sj=x1*ωj1+bj1+x2*ωj2+bj2+…x1*ωji+bji+…+xn*ωjn+bjn; yj = f(sj); Backpropagation is the process of error propagation from the output layer to the hidden layers and then back to the input layer. Gradient descent is used to adjust the connection weights and bias terms in the network to reduce the error between the network output and the desired output. The error function is the mean squared error, let d k For the desired output, o k For the actual output, the mean squared error is: The weight gradient error is calculated using the chain rule, and the weight adjustment formula for the connection parameters is: Where η is the learning rate; The output layer uses the sigmoid function as the activation function, with the output ranging from [0, 1]. The expression for the sigmoid function is:

3. The method for laying power optical cables according to claim 2, characterized in that: The training process of the BP neural network includes the following steps: S31: Determine the number of hidden layer nodes and adjust the weight matrix of the BP model; S32: The number of hidden layer nodes is determined, and the number of hidden layer neurons is used as an input parameter to test the same sample set data; The optimal number of neurons in the hidden layer can be determined by the following steps: S311: Testing from a small number of nodes; S312: Conduct network testing and training; S313: Test with increased node count; S314: Determine the number of nodes; During training, the network's generalization ability is increased, interference factors are reduced, and samples with the same number of iterations are selected for testing with different numbers of hidden nodes. The resulting error and maximum number of iterations are used as input parameters for further testing to obtain the model with the optimal number of hidden nodes.

4. The method for laying power optical cables according to claim 2, characterized in that: The initialization of the traction medium unwinding machine (3) in step S2 includes the following steps: S21: Initialize the network, assign random values ​​in the interval [-1, 1] to the connection parameter weights ω and the threshold b, and calculate the number of hidden layers p based on the number of input layers, output layers and adjustment constant a; S22: Training model assignment, providing the learning pattern pair [Xk, Yk] composed of the laying parameters of the power optical cable (4) to the neural network; S23: Forward propagation signal, calculate the net input and output of each unit in the hidden layer and output layer in sequence; S24: Backpropagation error, calculate the expected output dj and mean square error E, and calculate the generalized error of each unit in the output layer and hidden layer in turn; S25: Adjust the connection weights ω between the hidden layer and the output layer, and between the input layer and the hidden layer, as well as the threshold values ​​bji of each unit in the output layer; S26: Randomly provide the next optional learning data pair to the BP neural network, and return to step S23 until all learning data pairs have been trained; S27: Determine whether the global error of the network model can reach the required accuracy, i.e. whether E≤ε holds true. If it holds true, jump to step S28. If it does not hold true, return to step S23 if the number of learning iterations is ≤Nmax, otherwise continue to step S28. S28: Input the BP neural network with the laying length, material, diameter, ground and air markings, and traction medium quantification parameters of the power optical cable (4), and complete the normalization process; S29: The BP neural network outputs the corresponding optimal traction torque and traction speed, and restores the real data.

Images