Stacker hoist rope real-time monitoring method based on BP network
By BP network -based methods, real -time monitoring and evaluation of the stretching length of the stacking machine dragging rope solve the problem of the safety performance assessment of traction rope under long -term working conditions, and achieve higher monitoring accuracy and security, extend the extension The service life of the device.
Patent Information
- Application Number
- CN202411419153.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-12
- Publication Date
- 2025-01-24
AI Technical Summary
The prior art is difficult to monitor and evaluate the safety performance of the stacker traction rope in a long-term working state, resulting in wear, deformation and safety hazards of the traction rope.
Based on the BP network, the real -time monitoring system is established by obtaining the theoretical tensile length and actual stretching length of the traction rope, and the alarm and locking according to the safety factor and stretch length to ensure the safety of the traction rope.
The real -time safety performance assessment of the traction rope of the stack machine is realized, the accuracy and standardization of monitoring are improved, the wear and safety hazards of the traction rope are reduced, and the operating life of the stack machine is prolonged.
Smart Images
Figure CN119354718A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of stackers and relates to a real-time monitoring method for a traction rope of a stacker based on a BP network. Background Art
[0002] Automated high-bay warehouses are currently widely used due to their high automation, high operating efficiency and low storage costs. Among them, stackers shuttle and move in the lanes and automatically transport goods by receiving dispatch instructions from the warehouse control system WCS. They are the core equipment of intelligent high-bay warehouses. As an important part of the stacker, the stacker traction system is usually composed of a drive motor and a traction wire rope (hereinafter referred to as the traction rope), which has the characteristics of low cost, fast running speed, large load-bearing capacity and good bending flexibility.
[0003] At present, many scholars have conducted research on the elastic properties of traction ropes, mainly focusing on traction ropes in short-term working conditions, and are unable to conduct real-time monitoring and safety performance evaluation of traction ropes in long-term working conditions. During the operation of the stacker, the traction system is in a long-term working state, and the cargo platform needs to move up and down to position the goods at different heights. During this period, the traction rope has been in a taut state, and long-term use can easily cause wear, deformation, and even fatigue fracture. In addition, due to the strong elastic deformation capacity of the traction rope, when the cargo platform stops suddenly, the traction rope will stretch up and down, causing the cargo platform to vibrate longitudinally, which is difficult to quickly stabilize through the mechanical structure, causing the originally taut traction rope to further deform and wear, bringing safety hazards and affecting the operating life of the stacker.
[0004] Due to theoretical or technical reasons, many detection methods (electromagnetic, acoustic, optical, and ray methods, etc.) can only be studied in an experimental state. In actual projects, the performance evaluation of most traction ropes still relies on technical personnel to judge through visual inspection, which lacks accuracy and standardization, and cannot evaluate the safety of traction ropes in a long-term and timely manner. Summary of the invention
[0005] 1. Technical problems to be solved:
[0006] The current research mainly focuses on the traction rope in short-term working state, and it is impossible to conduct real-time monitoring and safety performance evaluation of the traction rope in long-term working state. The existing detection method can only be studied in the experimental state.
[0007] 2. Technical solution:
[0008] In order to solve the above problems, the present invention provides a real-time monitoring method for a stacker traction rope based on a BP network, comprising the following steps:
[0009] Step S01: Obtain the theoretical stretching length of the traction rope and the actual stretching length of the traction rope.
[0010] Step S02: within the allowable stretching range of normal operation [0, s1Δl), the system does not alarm and records the vibration waveform characteristics; if the actual stretching amount is in [s1Δl, s2Δl), the system issues a normal alarm, and the alarm can be manually eliminated after manual inspection and tensioning; if the actual traction rope stretching amount is greater than s2Δl, a danger alarm is issued and the system is locked, and the traction rope needs to be replaced before it can be unlocked, realizing the real-time monitoring function, where s1 and s2 are safety factors and Δl is the stretching length.
[0011] Furthermore, the method for obtaining the theoretical stretching length of the traction rope is as follows: the point where the traction rope is connected to the cargo platform is taken as the origin, vertically downward is the positive direction of X, the displacement generated by this point at time t is x(t), the mass of the traction rope is m1, the mass of the cargo platform and the cargo is m0, and m0>>m1, then the kinematic equation of the traction system is:
[0012]
[0013] Where c is the damping coefficient of the traction rope, k k is the elastic coefficient of the traction rope, then the solution is expressed as:
[0014]
[0015] Where ω d is the vibration angular frequency, f is the vibration frequency, is the amplitude, the constant A and the phase φ depend on the length of the free end of the traction rope and the speed of the cargo platform under the initial conditions. Since the system is an underdamped vibration system, 4(m0+m1)k k -c 2 >0, the overall system oscillates and decays.
[0016] The mass of the cargo platform and the cargo is measured in advance. The physical properties of the traction rope, such as the damping coefficient, are all fixed values. The speed of the cargo platform is directly read out through the PLC of the stacker. The mass m1 of the free end of the traction rope is obtained by multiplying the length of the free end of the traction rope by the linear density.
[0017] Furthermore, the method for obtaining the actual stretched length of the traction rope is:
[0018] Step S41: Install a vibration sensor on the crossbeam of the cargo platform connected to the towing rope.
[0019] Step S42: Use the stacker to load goods of different masses and control the loading platform to stop at the same position with a constant acceleration.
[0020] Step S43: When the stacker issues a stop command for the cargo platform, it starts to record the voltage signal of the vibration of the traction rope, and tries multiple times to obtain vibration waveforms corresponding to different load masses and different stretching amounts of the traction rope. At the same time, the output value of the vibration measuring instrument is recorded as the corresponding label value, and the output value is the actual maximum stretching amount of the traction rope.
[0021] Step S44: In an offline environment, the recorded voltage analog signal is imported into MATLAB, the signal characteristic value is identified using the peak-to-peak value method and imported into the BP neural network for learning.
[0022] Step S45: reference the Relu nonlinear function as the activation function f i , BP network hidden layer H k And the output layer Y j The outputs are:
[0023]
[0024] Where V i , W k is the weight, b i is the bias, X i is the input traction rope characteristic value, and the root mean square error RMSE is used as the loss function E p
[15] :
[0025]
[0026] In the formula, Y is the label value corresponding to the output of the vibration measuring instrument. The smaller the loss function, the higher the accuracy of the model. p The minimum value of is obtained, and the weights and biases are updated in reverse using the gradient descent method. The trained model is used to predict the test set samples to obtain the actual stretching length of the traction rope.
[0027] Furthermore, the data obtained by the vibration sensor transmits the vibration signal to the stacker PLC via RS485 communication or analog output, and transmits the signal to the host computer via Ethernet.
[0028] 3. Beneficial effects:
[0029] The present invention establishes a kinematic model of the longitudinal vibration of the traction rope, verifies the reliability of the model through finite element analysis and determines the theoretical stretching length of the traction rope; then uses the BP network to learn the waveform characteristics of the traction rope in actual work, thereby obtaining the actual elongation of the traction rope, thereby evaluating the safety of the traction rope, and finally builds a real-time monitoring system for the traction rope.
[0030] The present invention basically meets the actual engineering needs and greatly improves the standardization and accuracy of the safety performance evaluation of the traction rope during the actual operation of the stacker. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 It is a schematic diagram of the simplified model of the traction system.
[0032] Figure 2 It is the load distribution diagram of the traction rope.
[0033] Figure 3 is the deformation diagram of the traction rope.
[0034] Figure 4 It is the equivalent stress diagram of the traction rope.
[0035] Figure 5 It is an online monitoring framework for the elongation of the traction rope.
[0036] Figure 6 This is a comparison chart of the test set results.
[0037] Figure 7 It is a schematic diagram of the real-time monitoring system of the traction rope. DETAILED DESCRIPTION
[0038] The present invention is described in detail below with reference to the accompanying drawings and embodiments.
[0039] like Figure 7 As shown, the present invention provides a real-time monitoring method for a stacker traction rope based on a BP network, comprising the following steps:
[0040] Step S01: Obtain the theoretical stretching length of the traction rope and the actual stretching length of the traction rope.
[0041] Step S02: within the allowable stretching range of normal operation [0, s1Δl), the system does not alarm and records the vibration waveform characteristics; if the actual stretching amount is in [s1Δl, s2Δl), the system issues a normal alarm, and the alarm can be manually eliminated after manual inspection and tensioning; if the actual traction rope stretching amount is greater than s2Δl, a danger alarm is issued and the system is locked, and the traction rope needs to be replaced before it can be unlocked, realizing the real-time monitoring function, where s1 and s2 are safety factors and Δl is the stretching length.
[0042] The stacker traction system is mainly composed of motors, guide wheels, columns, cargo platforms and other components. During the operation of the lifting system, the vibration of the traction rope is mainly composed of lateral vibration and longitudinal vibration. Since the lateral vibration amplitude is small, it has little effect on the overall operation of the traction system and can be ignored. The longitudinal vibration characteristics of the traction rope are mainly considered. If the traction rope is regarded as a mass-spring-damper system with fixed parameters, the traction system can be simplified as follows: Figure 1As shown in the figure, since the length of the traction rope between the motor and the guide fixed pulley does not change during the entire working process, only the vibration characteristics of the free end of the traction rope escaping the guide fixed pulley are considered. And when the traction system is in a balanced state, the traction rope does not vibrate, so the vibration of the traction rope when the system stops and the cargo platform vibrates greatly is mainly considered.
[0043] In one embodiment, the method for obtaining the theoretical stretching length of the traction rope is: the point where the traction rope is connected to the cargo platform is taken as the origin, vertically downward is the positive direction of X, the displacement generated by this point at time t is x(t), the mass of the traction rope is m1, the mass of the cargo platform and the cargo is m0, and m0>>m1, then the kinematic equation of the traction system is:
[0044]
[0045] Where c is the damping coefficient of the traction rope, k k is the elastic coefficient of the traction rope, then the solution is expressed as:
[0046]
[0047] Where ω d is the vibration angular frequency, f is the vibration frequency, is the amplitude, the constant A and the phase φ depend on the length of the free end of the traction rope and the speed of the cargo platform under the initial conditions. Since the system is an underdamped vibration system, 4(m0+m1)k k -c 2 >0, the overall system oscillates and decays.
[0048] From formulas (2) and (3), it can be seen that the longitudinal vibration of the traction rope itself mainly depends on the mass of the stacker's cargo and the initial speed of the cargo platform. Substituting specific parameters into formula (2), the theoretical stretching length of the traction rope at a certain time and position can be determined. In engineering practice, the mass of the cargo platform and the cargo can be measured in advance, and the physical properties of the traction rope, such as the damping coefficient, are all fixed values. When the lifting system stops suddenly, the length of the free end of the traction rope and the speed of the cargo platform can be directly read through the stacker's PLC. The mass m1 of the free end of the traction rope can be obtained by multiplying the length of the free end of the traction rope by the line density.
[0049] In order to verify the accuracy of the traction rope vibration model (2), a finite element simulation analysis of the traction rope is performed. Taking a traction rope with a free end length of 20m as an example, the mass of the cargo platform is 500kg, the cargo is 1500kg, and the initial vertical descent speed of the cargo platform is 2m / s. The theoretical maximum tensile length of the traction rope is about 9.8192mm.
[0050] Due to the complex structure of the traction rope, in actual engineering it is usually equivalent to a cylinder with uniformly distributed mass. The cross-sectional area of the cylinder is equal to the sum of the cross-sectional areas of each steel wire of the traction rope, and the relative friction of each internal steel wire is ignored. Use CATIA to establish an equivalent model of the traction rope, and import the entire traction rope into ansys workbench for finite element analysis. Use hexahedral meshing for division, set the global unit size to 10mm, and obtain a total number of nodes of 377149 and a total number of units of 228326. After dividing the mesh, set the analysis conditions of the traction rope: The first step is to set the initial motion conditions of the traction rope, determine the initial velocity and apply tension Fz at both ends, such as Figure 2 As shown in the figure; the second step is to add constraints to one end of the traction rope to simulate the force when it stops. The result is as follows Figure 3 , 4 shown.
[0051] Figure 3 is the force deformation diagram of the traction rope, Figure 4 The equivalent stress diagram of the traction rope shows that if the traction rope is uniform in texture and there is no damage caused by shearing or bending inside, the stress is concentrated at the connection points at both ends. This means that under the tensile load, the danger point of the traction rope is near the restrained end. When the cargo platform stops suddenly, the traction rope is most likely to break first at the connection point with the cargo platform. The simulation result of the maximum tensile length of the traction rope is 9.5306mm, and the degree of agreement between the theoretical calculation and the simulation result [k] = 91.78%, which is within the error range and basically meets the accuracy requirements.
[0052] In order to evaluate the safety of the traction rope in real time, after determining the theoretical stretching length of the traction rope, it is necessary to obtain the actual stretching length of the traction rope. Figure 5 As shown, the method for obtaining the actual stretching amount of the traction rope includes the following steps: Step S41: Install a vibration sensor on the crossbeam of the cargo platform connecting the traction rope.
[0053] Step S42: Use the stacker to load goods of different masses and control the loading platform to stop at the same position with a constant acceleration.
[0054] Step S43: When the stacker issues a stop command for the cargo platform, it starts to record the voltage signal of the vibration of the traction rope, and tries multiple times to obtain vibration waveforms corresponding to different load masses and different stretching amounts of the traction rope. At the same time, the output value of the vibration measuring instrument is recorded as the corresponding label value, and the output value is the actual maximum stretching amount of the traction rope.
[0055] Step S44: In an offline environment, the recorded voltage analog signal is imported into MATLAB, the signal characteristic value is identified using the peak-to-peak value method and imported into the BP neural network for learning.
[0056] Step S45: reference the Relu nonlinear function as the activation function fi , BP network hidden layer H k And the output layer Y j The outputs are:
[0057]
[0058] Where V i , W k is the weight, b i is the bias, X i is the input traction rope characteristic value, and the root mean square error RMSE is used as the loss function E p
[15] :
[0059]
[0060] In the formula, Y is the label value corresponding to the output of the vibration measuring instrument. The smaller the loss function, the higher the accuracy of the model. p The minimum value of the rope is obtained by using the gradient descent method to reversely update the weights and biases, and the trained model is used to predict the test set samples to obtain the actual stretch length of the traction rope, such as Figure 6 shown.
[0061] Furthermore, the data obtained by the vibration sensor transmits the vibration signal to the stacker PLC via RS485 communication or analog output, and transmits the signal to the host computer via Ethernet.
[0062] In actual work, the real-time monitoring system of the traction rope is Figure 7 As shown in the figure, the monitoring system can meet the requirements by adding a vibration sensor and corresponding signal line to the original hardware of the stacker.
[0063] During normal operation of the stacker, the mass of the cargo and pallet is measured in advance by the weighing machine and stored in the host computer. It can be directly read when the stacker performs the task of storing and retrieving cargo. When the cargo platform is raised or lowered, the driving speed of the lifting servo motor and the free end length of the traction rope are controlled by the PLC and monitored by the rotary encoder of the lifting motor. The above parameters are passed to the host computer and substituted into formulas (2) and (3) to obtain the real-time theoretical elongation of the traction rope.
[0064] The vibration sensor is magnetically or bolted mounted on the upper side of the beam connecting the cargo platform to the traction rope. In one embodiment, the vibration signal is transmitted to the stacker PLC using RS485 communication or analog output, and the signal is transmitted to the host computer using Ethernet, thereby realizing real-time transmission and reception of the vibration signal.
Claims
1. A real-time monitoring method for a stacker traction rope based on a BP network, comprising the following steps: Step S01: obtaining the theoretical stretching length of the traction rope and the actual stretching length of the traction rope; Step S02: within the stretching range [0, s1Δl) allowed for normal operation, the system does not alarm and records the vibration waveform characteristics; If the actual stretching amount is in [s1Δl, s2Δl), the system will issue a normal alarm. After manual inspection and tensioning, the alarm can be manually eliminated. If the actual stretching amount of the traction rope is greater than s2Δl, a danger alarm will be issued and the system will be locked. The traction rope must be replaced before it can be unlocked to achieve real-time monitoring function, where s1 and s2 are safety factors and Δl is the stretching length.
2. The method for real-time monitoring of the traction rope of a stacker based on BP network according to claim 1, characterized in that: The method to obtain the theoretical stretching length of the traction rope is: take the point where the traction rope is connected to the cargo platform as the origin, vertically downward as the positive direction of X, the displacement generated by this point at time t is x(t), the mass of the traction rope is m1, the mass of the cargo platform and the cargo is m0, and m0>>m1, then the kinematic equation of the traction system is: Where c is the damping coefficient of the traction rope, k k is the elastic coefficient of the traction rope, then the solution is expressed as: Where ω d is the vibration angular frequency, f is the vibration frequency, d is the amplitude, the constant A and the phase φ depend on the length of the free end of the traction rope and the speed of the cargo platform under the initial conditions. Since the system is an underdamped vibration system, 4(m0+m1)k k -c 2 >0, the overall system oscillates and decays.
3. The real-time monitoring method for stacker traction rope based on BP network as claimed in claim 2 is characterized in that: The mass of the cargo platform and the cargo is measured in advance. The physical properties of the traction rope, such as the damping coefficient, are all fixed values. The speed of the cargo platform is directly read out through the PLC of the stacker. The mass m1 of the free end of the traction rope is obtained by multiplying the length of the free end of the traction rope by the linear density.
4. The method for real-time monitoring of the traction rope of a stacker based on BP network according to claim 1, characterized in that: The method to obtain the actual stretching length of the traction rope is: Step S41: installing a vibration sensor on the crossbeam of the cargo platform connected to the traction rope; Step S42: using a stacker to load goods of different masses and controlling the loading platform to stop at the same position with a constant acceleration; Step S43: When the stacker issues a stop command for the cargo platform, it starts to record the voltage signal of the vibration of the traction rope, and tries several times to obtain the vibration waveforms of the traction rope with different load masses and different stretching amounts, and records the output value of the vibration measuring instrument as the corresponding label value, and the output value is the actual maximum stretching amount of the traction rope; Step S44: in an offline environment, the recorded voltage analog signal is imported into MATLAB, the signal characteristic value is identified by using the peak-to-peak value method and imported into the BP neural network for learning; Step S45: reference the Relu nonlinear function as the activation function f i , BP network hidden layer H k And the output layer Y j The outputs are: Where V i , W k is the weight, b i is the bias, X i is the input traction rope characteristic value, and the root mean square error RMSE is used as the loss function E p [15]: In the formula, Y is the label value corresponding to the output of the vibration measuring instrument. The smaller the loss function, the higher the accuracy of the model. p The minimum value of is obtained, and the weights and biases are updated in reverse using the gradient descent method. The trained model is used to predict the test set samples to obtain the actual stretching length of the traction rope.
5. The real-time monitoring method for stacker traction rope based on BP network as claimed in claim 4, characterized in that; The data obtained by the vibration sensor transmits the vibration signal to the stacker PLC via RS485 communication or analog output, and transmits the signal to the host computer via Ethernet.