Intelligent control method and system for dam filler rolling process
Through the numerical simulation model and agent model of dam filler rolling, combined with the particle swarm algorithm, a compaction quality evaluation and incremental prediction model is built, which solves the problem of crushing parameters optimization, realizes intelligent control, improves compaction quality and efficiency, and reduces costs.
Patent Information
- Application Number
- CN202510865363.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-26
- Publication Date
- 2025-07-25
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The lack of effective crushing parameter optimization methods and continuous compaction quality detection methods under variable crushing parameter conditions in the prior art has resulted in the failure to fully realize intelligent compaction technology, and the on-site test cost is high and time-consuming, making it difficult to achieve real-time optimization.
The numerical simulation model of dam filler rolling is adopted, combined with the agent model and particle swarm algorithm, and the compaction quality evaluation and incremental prediction model are constructed, and the crushing parameters are detected and optimized in real time, and intelligent control is achieved through the signal acquisition system and the electronic control system.
Real-time detection and optimization of compaction quality under variable rolling parameters is achieved, the packing compaction quality and efficiency are improved, and cost and time consumption are reduced.
Smart Images

Figure CN120370823A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent construction of hydropower projects, and particularly relates to an intelligent control method and system for the rolling process of dam fillers. Background Art
[0002] Filling and rolling is a key link in dam construction. The compaction quality and efficiency of fillers are directly related to the dam safety, construction period and investment. Traditional rolling uses fixed rolling parameters, and the compaction quality is detected by post-sampling. There is no active control over the rolling process. According to the vibration compaction theory, the influence of vibration compaction parameters on the compaction effect is different in different compaction stages. Optimizing the vibration compaction parameters according to the compaction state can improve the compaction effect and efficiency.
[0003] To overcome the shortcomings of traditional rolling and improve the efficiency and quality of compaction operations, some scholars and manufacturers have proposed continuous compaction quality detection technology. By analyzing the time and frequency domains of vibration acceleration, continuous compaction quality detection indicators are proposed to conduct real-time detection and evaluation of the compaction quality of fillers. Currently, this technology has been applied in actual projects. Further, intelligent compaction technology has been proposed. Based on the continuous compaction quality detection, this technology adds a feedback control function for rolling working parameters to optimize the control of rolling parameters according to the compaction state of fillers.
[0004] However, although the concept of intelligent compaction has been proposed for many years, intelligent compaction has not been fully realized yet. The main reasons are: first, there is a lack of effective and feasible rolling parameter optimization methods; second, the current continuous compaction quality detection technology is applicable to the condition of fixed rolling parameters. Changing the vibration rolling parameters will have a significant impact on the dynamic response of the vibratory roller-filler system, and thus have a significant impact on the continuous compaction quality detection indicator values of fillers. To truly achieve intelligent compaction, it is necessary to propose effective and feasible rolling parameter optimization methods and continuous compaction quality detection and evaluation methods under variable rolling parameter conditions.
[0005] Meanwhile, the current continuous detection of compaction quality mainly relies on on-site tests to establish the relationship between compaction quality and continuous detection indicators. However, there is a highly non-linear relationship among rolling parameters, the compaction state of the filler, and the dynamic response of the vibratory roller-filler system. To explore the influence of different rolling parameters on the compaction effect of fillers in different compaction states, as well as the influence of different rolling parameters and filler compaction states on the continuous detection results, a large number of tests need to be carried out. The economic and time costs of on-site tests are huge and difficult to achieve. With the maturity of numerical simulation methods, using numerical simulation to simulate the rolling process of the roller and analyze the influence of rolling parameters on the compaction effect and the dynamic response of the vibratory roller-filler system has become a feasible method. Compared with on-site tests, the numerical simulation method does not require a large amount of site, mechanical equipment, and test personnel, saving a lot of manpower and material resources, and shortening the time-consuming. It can analyze more working conditions. However, numerical simulation is still time-consuming. For example, one calculation may take several hours and still cannot be used for the real-time optimization of the actual rolling process. As an efficient data-driven mathematical modeling method, the surrogate model can quickly establish the mapping relationship between the input and output of a multi-dimensional variable system using a small amount of sample data, reducing time and computational costs, and providing a solution for real-time evaluation of the compaction quality of fillers and optimization of rolling parameters. Summary of the Invention
[0006] Aiming at the defects existing in the prior art, the present invention provides an intelligent control method and system for the rolling process of dam fillers, which can effectively solve the above problems.
[0007] The technical solution adopted by the present invention is as follows:
[0008] The present invention provides an intelligent control method for the rolling process of dam fillers, including the following steps:
[0009] Step 1: Establish a numerical simulation model for the rolling of dam fillers;
[0010] Step 2: Conduct indoor tests and on-site rolling tests to calibrate the parameters of the numerical simulation model and obtain the calibrated numerical simulation model;
[0011] Step 3: Use the calibrated numerical simulation model to carry out multi-condition filler rolling simulation calculations and obtain simulation calculation results;
[0012] Step 4: Use the simulation calculation results as the training set to construct a compaction quality evaluation surrogate model and a compaction quality increment prediction surrogate model respectively;
[0013] Step 5: Based on the compaction quality evaluation surrogate model and the compaction quality increment prediction surrogate model respectively, establish a compaction quality evaluation system and a rolling parameter optimization system;
[0014] Step 6: During the rolling process, the signal acquisition system collects the dynamic responses of the vibratory roller - filler in real time and inputs them into the compaction quality assessment system;
[0015] Step 7: The compaction quality assessment system evaluates the current compaction quality of the filler and transmits it to the rolling parameter optimization system;
[0016] Step 8: Based on the current compaction quality of the filler, the rolling parameter optimization system takes the maximum compaction efficiency as the optimization goal and outputs the optimal rolling parameters for the next pass of rolling;
[0017] Step 9: The electronic control system of the vibratory roller adjusts the rolling parameters to the optimal rolling parameters and conducts the next pass of rolling.
[0018] Preferably, the numerical simulation model for the rolling of the dam filler is a two - dimensional or three - dimensional numerical simulation model; the length of the rolling area along the rolling direction is at least 20 m, the thickness of the filler is at least 4 m, and the filler is stratified according to the actual paving thickness; the filler uses solid elements with an element size not greater than 10 cm; the filler uses an elastoplastic constitutive model; the vibratory roller is modeled as a whole or only the vibration wheel model is established, and the vibration wheel is set as a rigid body; infinite element cells are set around and at the bottom of the model to prevent the reflection of vibration waves at the boundary from affecting the calculation results.
[0019] Preferably, the indoor tests include static and dynamic triaxial tests, shear tests, and vibration compaction tests of the filler, which are used to obtain various physical and mechanical parameters required for the filler in the numerical simulation model.
[0020] Preferably, the on - site rolling tests select typical working conditions to conduct multiple groups of on - site rolling tests. During the rolling process, the acceleration of the vibration wheel and the interaction force between the vibration wheel and the filler are detected, and the density and settlement deformation of the filler are measured.
[0021] Preferably, calibrate the parameters of the numerical simulation model, specifically:
[0022] Adjust the parameters of the numerical simulation model according to the on - site test results so that the error between the simulation calculation results of the numerical simulation model and the on - site measured results is within 10%.
[0023] Preferably, conduct the simulation calculations of the rolling of the filler under multiple working conditions, specifically:
[0024] Set different levels for the rolling parameters and the compaction state of the filler, and adopt orthogonal sampling, uniform sampling, random sampling, and Latin hypercube sampling to design a multi - working - condition test scheme and conduct simulation calculations, and extract the calculation results of the acceleration of the vibration wheel, the interaction force between the vibration wheel and the filler, and the compaction quality under different rolling working conditions.
[0025] Preferably, for the compaction quality evaluation proxy model, the model inputs are the vibration wheel acceleration eigenvalue, the vibration wheel-packing interaction force, the rolling vehicle speed, the vibration frequency, and the excitation force, and the model output is the current packing compaction quality. The model is as follows:
[0026]
[0027] Where: is the packing compaction quality after the i-th pass of rolling; A i is the vibration wheel acceleration eigenvalue of the i-th pass of rolling; F si is the vibration wheel-packing interaction force of the i-th pass of rolling; v i is the rolling vehicle speed of the i-th pass of rolling; f i is the vibration frequency of the i-th pass of rolling; F ei is the excitation force of the i-th pass of rolling;
[0028] For the compaction quality increment prediction proxy model, the inputs are the current packing compaction quality after the i-th pass of rolling, the rolling vehicle speed, the vibration frequency, and the excitation force used in the next (i + 1)-th pass of rolling, and the output is the compaction quality increment of the packing in the (i + 1)-th pass of rolling. The model is as follows:
[0029]
[0030] Where: is the compaction quality increment of the packing after the (i + 1)-th pass of rolling; is the rolling vehicle speed of the (i + 1)-th pass of rolling; is the vibration frequency of the (i + 1)-th pass of rolling; is the excitation force of the (i + 1)-th pass of rolling.
[0031] Preferably, the compaction quality evaluation system is specifically used for:
[0032] During the rolling process, the signal acquisition system collects the original acquisition data of the dynamic response of the vibratory roller-packing in real time and performs preprocessing to obtain the preprocessed data, including the vibration wheel acceleration eigenvalue, the vibration wheel-packing interaction force, the rolling vehicle speed, the vibration frequency, and the excitation force; and inputs the preprocessed data into the compaction quality evaluation proxy model to output the current packing compaction quality.
[0033] Preferably, step 8 is specifically:
[0034] The rolling parameter optimization system aims to maximize the compaction quality increment efficiency per unit packing, that is:
[0035]
[0036] Where: is the compaction efficiency of the (i + 1)-th pass of rolling; is the increment of the compaction quality of the filler after the (i + 1)-th pass of rolling, calculated by the compaction quality increment prediction proxy model; is the rolling speed of the (i + 1)-th pass of rolling;
[0037] The constraint conditions include:
[0038] (1) The constraint of no chaotic vibration;
[0039] Perform a fast Fourier transform on the eigenvalue A of the vibratory drum acceleration in the i-th pass of rolling i , then perform spectral analysis, and extract the amplitudes A 1 / 2 , A 3 / 2 , A 5 / 2 , A 7 / 2 ... of the 1 / 2f0, 3 / 2f0, 5 / 2f0, 7 / 2f0... sub-harmonics, where f0 is the frequency; accumulate the amplitudes of the sub-harmonics, and the accumulated value is not greater than the set value , that is, satisfy the following formula:
[0040]
[0041] (2) The vibration frequency, rolling speed, and exciting force are within the adjustment range of the vibratory roller, that is, satisfy the following formula:
[0042]
[0043] Where: f min , v min and F emin are the minimum vibration frequency, minimum rolling speed, and minimum exciting force of the roller respectively; f max , v max and F emax are the maximum vibration frequency, maximum rolling speed, and maximum exciting force of the roller respectively;
[0044] (3) The constraint of meeting the compaction uniformity requirement:
[0045] The horizontal displacement of the roller within one vibration period, that is, the impact spacing, needs to meet the following requirements:
[0046]
[0047] Where: is the rolling speed of the (i + 1)-th pass of rolling; is the vibration frequency of the (i + 1)-th pass of rolling; is the minimum value of the impact spacing; is the maximum value of the impact spacing;
[0048] The optimization of the rolling parameters is realized by the particle swarm algorithm, including:
[0049] S1. Initialize the parameters of the particle swarm, including the particle swarm size, particle dimension, number of iterations, inertia weight, learning factors, and iteration step size;
[0050] S2. Randomly initialize the position and velocity of each particle;
[0051] S3. Calculate the fitness of each particle using the following formula;
[0052]
[0053] where: is the fitness; is the increment of the compacted quality of the filler after the next (i.e., the (i + 1)-th pass) compaction, calculated by the compaction quality increment prediction surrogate model; is the compaction vehicle speed for the next (i.e., the (i + 1)-th pass) compaction;
[0054] S4. Update the historical best fitness and position of each particle and the population;
[0055] S5. Update the particle velocity and position;
[0056]
[0057] where: is the velocity of the i-th particle at the (k + 1)-th iteration; is the weight coefficient; is the velocity of the i-th particle at the k-th iteration; and are the learning factors; and are two random numbers from a normal distribution, in the range [0, 1]; is the solution corresponding to the best fitness obtained by the i-th particle up to the k-th iteration; is the solution corresponding to the best fitness among all particles up to the k-th iteration; and are the solutions corresponding to all particles at the k-th and (k + 1)-th iterations, respectively;
[0058] Update the individual position when the new particle position satisfies the constraint conditions; do not update the individual position when the new particle position does not satisfy the constraint conditions;
[0059] S6. Determine whether the end condition is satisfied;
[0060] S7. If satisfied, output the optimal solution; if not, return to step S3 to repeat the process.
[0061] The present invention also provides an intelligent control system for the rolling process of dam fillers, which includes a signal acquisition system, a compaction quality evaluation system, a rolling parameter optimization system, and an electric control system;
[0062] The signal acquisition system, through a satellite customization device, an acceleration sensing device, and a vibrating wheel-filler interaction detection device installed on a vibratory roller, real-time collects the position information of the roller, the vibrating wheel acceleration signal, and the vibrating wheel-filler interaction information during the vibratory rolling process, and inputs them into the compaction quality evaluation system;
[0063] The compaction quality evaluation system preprocesses the data collected by the signal acquisition system to obtain the vibrating wheel acceleration characteristic value, the vibrating wheel-filler interaction force, the rolling vehicle speed, the vibration frequency, and the excitation force, and then inputs them into the compaction quality evaluation agent model to output the current filler compaction quality; the current filler compaction quality is input into the rolling parameter optimization system;
[0064] The rolling parameter optimization system, based on the current compaction quality of the filler, aims to maximize the compaction efficiency of the next pass of rolling, and uses the particle swarm algorithm to optimize the rolling parameters of the next pass. Among them, the compaction quality increment prediction agent model is used to calculate the particle fitness, and the optimal rolling parameters calculated by the rolling parameter optimization system are input into the electric control system;
[0065] The electric control system automatically adjusts the mechanical and hydraulic systems of the roller according to the obtained optimal rolling parameters, and controls the rolling vehicle speed, vibration frequency, and excitation force to the optimal rolling parameters.
[0066] An intelligent control method and system for the rolling process of dam fillers provided by the present invention have the following advantages:
[0067] An intelligent control method and system for the rolling process of dam fillers provided by the present invention are applicable to the rolling of dam fillers, and can also be extended to earth and stone filling projects such as highways, railways, and airports, which can effectively improve the compaction quality and efficiency of fillers. Description of the Drawings
[0068] Figure 1 It is a flowchart of the intelligent control method for the rolling process of dam fillers provided by the present invention;
[0069] Figure 2 It is a schematic diagram of the semi-harmonic amplitude accumulation calculation provided by the present invention;
[0070] Figure 3 It is a flowchart of the rolling parameter optimization based on the particle swarm algorithm provided by the present invention;
[0071] Figure 4 It is a composition diagram of the intelligent control system for the rolling process of dam fillers provided by the present invention. Detailed Embodiments
[0072] In order to make the technical problems, technical solutions and beneficial effects solved by the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0073] The present invention relates to an intelligent control method and system for the rolling process of dam fillers, belonging to the field of water conservancy and hydropower engineering construction. Specifically, it is an intelligent control method and system for the rolling process of dam fillers based on the surrogate model and particle swarm algorithm. This method is applicable to the rolling of dam fillers and can also be extended to earth and stone filling projects such as highways, railways, and airports, which can effectively improve the compaction quality and efficiency of fillers.
[0074] Refer to Figure 1 , the present invention provides an intelligent control method for the rolling process of dam fillers, including the following steps:
[0075] Step 1: Establish a numerical simulation model for the rolling of dam fillers;
[0076] Step 2: Conduct indoor tests and on-site rolling tests, calibrate the parameters of the numerical simulation model, and obtain the calibrated numerical simulation model;
[0077] Step 3: Use the calibrated numerical simulation model to conduct multi-condition filler rolling simulation calculations and obtain simulation calculation results;
[0078] Step 4: Take the simulation calculation results as the training set, and construct a compaction quality evaluation surrogate model and a compaction quality increment prediction surrogate model respectively;
[0079] Step 5: Based on the compaction quality evaluation surrogate model and the compaction quality increment prediction surrogate model respectively, establish a compaction quality evaluation system and a rolling parameter optimization system;
[0080] Step 6: During the rolling process, the signal acquisition system real-time collects the dynamic response of the vibratory roller - filler and inputs it into the compaction quality evaluation system;
[0081] Step 7: The compaction quality evaluation system evaluates the current compaction quality of the filler and transmits it to the rolling parameter optimization system;
[0082] Step 8: The rolling parameter optimization system takes the current compaction quality of the filler as the optimization target with the maximum compaction efficiency, and outputs the optimal rolling parameters for the next pass of rolling;
[0083] Step 9: The electronic control system of the vibratory roller adjusts the rolling parameters to the optimal rolling parameters and conducts the next pass of rolling.
[0084] The following is a detailed introduction to each step:
[0085] Step 1: Establish a numerical simulation model for the compaction of dam fill materials;
[0086] The numerical simulation model for the compaction of dam fill materials can be a two-dimensional or three-dimensional numerical simulation model; the length of the compaction area along the compaction direction is at least 20 m, the thickness of the fill material is at least 4 m, and the fill material is stratified according to the actual paving thickness; the fill material uses solid elements with an element size not greater than 10 cm; the fill material uses elastoplastic constitutive models such as the Drucker-Prager model and the hypoplastic model; the vibratory roller can be modeled as a whole or only the vibration drum model can be established, and the vibration drum is set as a rigid body; infinite elements are set around and at the bottom of the model to prevent the reflection of vibration waves at the boundary from affecting the calculation results.
[0087] Step 2: Conduct indoor tests and on-site compaction tests to calibrate the parameters of the numerical simulation model and obtain the calibrated numerical simulation model;
[0088] The indoor tests may include relevant tests such as static and dynamic triaxial tests, shear tests, and vibration compaction tests of the fill material, which are used to obtain various physical and mechanical parameters required for the fill material in the numerical simulation model.
[0089] For the on-site compaction test, select typical working conditions for multiple groups of on-site compaction tests. During the compaction process, measure the acceleration of the vibration drum and the interaction force between the vibration drum and the fill material, and measure the fill material density and settlement deformation.
[0090] Step 3: Use the calibrated numerical simulation model to conduct numerical simulation calculations for the compaction of fill materials under multiple working conditions and obtain the simulation calculation results;
[0091] Calibrating the parameters of the numerical simulation model specifically means:
[0092] Adjust the parameters of the numerical simulation model according to the on-site test results so that the error between the simulation calculation results of the numerical simulation model and the on-site measured results is within 10%.
[0093] The conduct of numerical simulation calculations for the compaction of fill materials under multiple working conditions specifically means:
[0094] Set different levels for the compaction parameters (compaction speed, vibration frequency, excitation force) and the compaction state of the fill material within the possible range. Orthogonal sampling, uniform sampling, random sampling, and Latin hypercube sampling and other methods can be used to design a multi-condition test plan and conduct simulation calculations, and extract calculation results such as the acceleration of the vibration drum, the interaction force between the vibration drum and the fill material, and the compaction quality (fill material density, fill material compaction deformation) under different compaction conditions.
[0095] Step 4: Use the simulation calculation results as the training set to construct a compaction quality evaluation surrogate model and a compaction quality increment prediction surrogate model respectively;
[0096] Specifically, 80% of the calculation results of the numerical simulation model are used as the model training set, and 20% are used as the model validation set. Specific methods such as the Kriging model, artificial neural network, radial basis function, support vector regression model, polynomial regression model, etc. can be used to construct the surrogate model. The error between the calculation results of the surrogate model and the numerical simulation calculation results is within 5%. After each batch of rolling operations in the actual project, the on-site measured data is added to the model training set to update the surrogate model.
[0097] For the surrogate model for predicting the compaction quality increment, the input is the compaction quality of the filler after the current i-th pass of rolling, the rolling vehicle speed, vibration frequency, and excitation force used in the next (i + 1)-th pass of rolling, and the output is the compaction quality increment of the filler in the (i + 1)-th pass of rolling. The model is as follows:
[0098]
[0099] Where: is the compaction quality increment of the filler after the (i + 1)-th pass of rolling; is the rolling vehicle speed in the (i + 1)-th pass of rolling; is the vibration frequency in the (i + 1)-th pass of rolling; is the excitation force in the (i + 1)-th pass of rolling.
[0100] Step 5: Based on the compaction quality evaluation surrogate model and the compaction quality increment prediction surrogate model respectively, establish a compaction quality evaluation system and a rolling parameter optimization system;
[0101] The compaction quality evaluation system is specifically used for:
[0102] During the rolling process, the signal acquisition system real-time collects the original acquisition data of the dynamic response of the vibratory roller - filler, and performs preprocessing to obtain the preprocessed data, including the vibration wheel acceleration eigenvalue, the interaction force between the vibration wheel and the filler, the rolling vehicle speed, the vibration frequency, and the excitation force; input the preprocessed data into the compaction quality evaluation surrogate model, and output the current compaction quality of the filler.
[0103] Step 6: During the rolling process, the signal acquisition system real-time collects the dynamic response of the vibratory roller - filler and inputs it into the compaction quality evaluation system;
[0104] Step 7: The compaction quality evaluation system evaluates the current compaction quality of the filler and transmits it to the rolling parameter optimization system;
[0105] Specifically, for the compaction quality evaluation surrogate model, the model input is the vibration wheel acceleration eigenvalue, the interaction force between the vibration wheel and the filler, the rolling vehicle speed, the vibration frequency, and the excitation force, and the model output is the current compaction quality of the filler. The model is as follows:
[0106]
[0107] Wherein: is the compaction quality of the filler after the i-th pass of rolling; A i is the eigenvalue of the vibratory drum acceleration during the i-th pass of rolling; F si is the interaction force between the vibratory drum and the filler during the i-th pass of rolling; v i is the rolling vehicle speed during the i-th pass of rolling; f i is the vibration frequency during the i-th pass of rolling; F ei is the exciting force during the i-th pass of rolling;
[0108] Step 8: The rolling parameter optimization system outputs the optimal rolling parameters for the next pass of rolling with the maximum compaction efficiency as the optimization goal according to the current compaction quality of the filler;
[0109] The rolling parameter optimization system aims to maximize the incremental efficiency of the compaction quality per unit of filler, that is:
[0110]
[0111] Wherein: is the compaction efficiency of the (i + 1)-th pass of rolling; is the increment of the compaction quality of the filler after the (i + 1)-th pass of rolling, calculated by the compaction quality increment prediction proxy model; is the rolling vehicle speed of the (i + 1)-th pass of rolling;
[0112] The optimization of rolling parameters needs to meet the constraint conditions, and the constraint conditions include: (1) the constraint of no chaotic vibration; (2) the vibration frequency, rolling vehicle speed, and exciting force are within the adjustment range of the vibratory roller; (3) meet the requirement of compaction uniformity.
[0113] (1) The constraint of no chaotic vibration;
[0114] The constraint of no chaotic vibration means that according to the dynamic analysis, when the vibration frequency of the roller is close to the natural frequency of the vibratory roller - filler system, resonance will occur, resulting in too large an amplitude, generating chaotic vibration, damaging mechanical equipment, and endangering the physical and mental health of the driver. During chaotic vibration, a large semi-harmonic peak will appear in the acceleration spectrum of the vibratory drum. Based on this, the following method is used to judge no chaotic vibration:
[0115] Perform a fast Fourier transform (FFT) on the eigenvalue A i of the vibratory drum acceleration during the i-th pass of rolling, and then perform spectrum analysis. As Figure 2 shown, extract the amplitudes A 1 / 2 、A 3 / 2 、A5 / 2 、A 7 / 2 ···, where f0 is the frequency; the amplitudes of the sub-harmonics are accumulated, and the accumulated value is not greater than the set value , that is, the following formula is satisfied:
[0116]
[0117] (2) The vibration frequency, rolling speed, and excitation force are within the adjustment range of the vibratory roller, that is, the following formula is satisfied:
[0118]
[0119] Where: f min , v min and F emin are the minimum vibration frequency, minimum rolling speed, and minimum excitation force of the roller respectively; f max , v max and F emax are the maximum vibration frequency, maximum rolling speed, and maximum excitation force of the roller respectively;
[0120] (3) Meet the constraint of the compaction uniformity requirement:
[0121] The horizontal displacement of the roller within one vibration period, that is, the impact spacing, has an impact on the compaction uniformity. If the impact spacing is too large, it will cause ripple phenomena on the compacted surface; if the impact spacing is too small, it will cause over-compaction, resulting in a reduction in compaction quality. The impact spacing needs to meet the following requirements:
[0122]
[0123] Where: is the rolling speed of the (i + 1)-th pass of rolling; is the vibration frequency of the (i + 1)-th pass of rolling; is the minimum value of the impact spacing, which can be taken as 0.02 m; is the maximum value of the impact spacing, which can be taken as 0.03 m;
[0124] The optimization of the rolling parameters is realized by the Particle Swarm Optimization (PSO) algorithm, and the process is as Figure 3 shown:
[0125] S1, Initialize the parameters of the particle swarm, including the particle swarm size, particle dimension, number of iterations, inertia weight, learning factor, iteration step size, etc.;
[0126] S2, Randomly initialize the position and velocity of each particle;
[0127] S3, Calculate the fitness of each particle using the following formula;
[0128]
[0129] Wherein: is the fitness; is the increment of the compaction quality of the filler after the next pass, i.e., the (i + 1)-th pass, which is calculated by the compaction quality increment prediction proxy model; is the rolling vehicle speed of the next pass, i.e., the (i + 1)-th pass;
[0130] S4. Update the historical optimal fitness and position of each particle and the population;
[0131] S5. Update the particle velocity and position according to the following formula;
[0132]
[0133] Wherein: is the velocity of the i-th particle at the (k + 1)-th iteration; is the weight coefficient; is the velocity of the i-th particle at the k-th iteration; and are the learning factors; and are two random numbers taking values from a normal distribution, with the range of [0, 1]; is the solution corresponding to the optimal fitness obtained by the i-th particle up to the k-th iteration; is the solution corresponding to the optimal fitness obtained among all particles up to the k-th iteration; and are the solutions corresponding to all particles at the k-th and (k + 1)-th iterations respectively;
[0134] Update the individual position when the new position of the particle satisfies the constraint conditions; do not update the individual position when the new position of the particle does not satisfy the constraint conditions;
[0135] S6. Determine whether the end condition is satisfied;
[0136] S7. If satisfied, output the optimal solution; if not satisfied, return to step S3 to repeat the process.
[0137] Step 9: The electronic control system of the vibratory roller adjusts the rolling parameters to the optimal rolling parameters and performs the next pass of rolling.
[0138] To implement the intelligent control method for the dam rolling process, the present invention also provides an intelligent control system for the dam filler rolling process, as Figure 4 shown, which includes a signal acquisition system, a compaction quality evaluation system, a rolling parameter optimization system, and an electronic control system;
[0139] The signal acquisition system collects the position information of the roller, the acceleration signal of the vibrating wheel, and the interaction information between the vibrating wheel and the filler during the vibrating compaction process in real time through a satellite customization device, an acceleration sensing device, a vibrating wheel-filler interaction detection device, etc. installed on the vibrating roller, and inputs them into the compaction quality evaluation system. Among them, the acquisition frequencies of the acceleration sensing device and the vibrating wheel-filler interaction force detection device are not less than 1000 Hz.
[0140] The compaction quality evaluation system preprocesses the data collected by the signal acquisition system, including acceleration filtering, FFT analysis, integration, calculation of the roller position and vehicle speed, and calculation of the interaction force between the vibrating wheel and the filler, to obtain characteristic values of the vibrating wheel acceleration, the interaction force between the vibrating wheel and the filler, the compaction vehicle speed, the vibration frequency, the exciting force, etc., and then inputs them into the compaction quality evaluation agent model to output the current compaction quality of the dam filler; the current compaction quality of the filler is input into the rolling parameter optimization system.
[0141] The rolling parameter optimization system optimizes the rolling parameters for the next pass based on the current compaction quality of the filler with the goal of maximizing the compaction efficiency of the next pass. Among them, the compaction quality increment prediction agent model is used to calculate the particle fitness, and the optimal rolling parameters calculated by the rolling parameter optimization system are input into the electronic control system.
[0142] The electronic control system can automatically adjust the mechanical and hydraulic systems of the roller according to the obtained optimal rolling parameters, and control the rolling vehicle speed, vibration frequency, and exciting force to the optimal rolling parameters.
[0143] Compared with the traditional fixed compaction scheme, the use of the intelligent compaction control method and system can effectively improve the compaction quality and compaction efficiency.
[0144] Different from other patent applications or research papers, the biggest feature of the present invention is: using a method combining an agent model and a particle swarm algorithm to optimize the rolling parameters. Taking the high-precision numerical simulation results of dam filler compaction as the training set, a compaction quality evaluation agent model and a compaction quality increment prediction agent model are constructed to realize the real-time detection and evaluation of compaction quality under variable rolling parameter conditions. Taking the real-time detected and evaluated compaction quality as the input, with the goal of maximizing the efficiency of the compaction quality increment per unit of the next pass, the rolling parameters for the next pass are optimized based on the particle swarm algorithm, ensuring the compaction quality while improving the compaction efficiency.
[0145] The present invention proposes an intelligent control method and system for the rolling process of dam fillers, which is applicable to the intelligent compaction control of dam fillers and can also be used in earthwork filling projects such as highways, railways, and airports. The core protection points include:
[0146] (1) Generate data through numerical simulation of dam filler compaction, construct a compaction quality evaluation surrogate model and a compaction quality increment prediction surrogate model, and realize real-time detection and evaluation of compaction quality and prediction of compaction quality increment under different compaction parameter conditions during the intelligent compaction control process.
[0147] (2) Propose an optimization goal of maximizing the efficiency of the compaction quality increment per unit compaction in the next pass, and propose specific optimization functions and constraints, and propose an implementation method based on the particle swarm algorithm.
[0148] (3) Propose an intelligent control system for the dam filler compaction process, which consists of four parts: a signal acquisition system, a compaction quality evaluation system, a compaction parameter optimization system, and an electric control system, and propose the main components and functions of each system.
[0149] The present invention uses the numerical simulation results to train a surrogate model, realizing the compaction quality evaluation under variable compaction parameter conditions and the rapid analysis of the influence of different compaction parameters on the compaction effect. The optimization goal of maximizing the efficiency of the compaction quality increment per unit compaction in the next pass of the present invention can improve the compaction efficiency; the particle swarm algorithm is used for optimization, which is suitable for high-dimensional and nonlinear problems, has strong global search ability, and can accurately obtain the optimal compaction parameters.
[0150] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.
Claims
1. An intelligent control method for the rolling process of dam filler, characterized in that It includes the following steps: Step 1: Establish a numerical simulation model for the compaction of dam fill materials; Step 2: Conduct indoor tests and on-site compaction tests to calibrate the parameters of the numerical simulation model and obtain the calibrated numerical simulation model; Step 3: Use the calibrated numerical simulation model to carry out multi-condition fill compaction simulation calculations and obtain the simulation calculation results; Step 4: Use the simulation calculation results as the training set to construct a compaction quality evaluation surrogate model and a compaction quality increment prediction surrogate model respectively; Step 5: Based on the compaction quality evaluation surrogate model and the compaction quality increment prediction surrogate model respectively, establish a compaction quality evaluation system and a compaction parameter optimization system; Step 6: During the compaction process, the signal acquisition system real-time collects the dynamic response of the vibratory roller-fill material and inputs it into the compaction quality evaluation system; Step 7: The compaction quality evaluation system evaluates the current fill compaction quality and transmits it to the compaction parameter optimization system; Step 8: The compaction parameter optimization system, based on the current fill compaction quality and with the maximum compaction efficiency as the optimization goal, outputs the optimal compaction parameters for the next pass of compaction; Step 9: The electronic control system of the vibratory roller adjusts the compaction parameters to the optimal compaction parameters and conducts the next pass of compaction.
2. The intelligent control method for the rolling process of dam filling materials according to claim 1, wherein, The numerical simulation model for the compaction of dam fill materials is a two-dimensional or three-dimensional numerical simulation model; the length of the compaction area along the compaction direction is at least 20 m, the fill thickness is at least 4 m, and the fill is stratified according to the actual paving thickness; the fill uses solid elements with a unit size not greater than 10 cm; the fill uses an elastoplastic constitutive model; the vibratory roller is modeled as a whole or only the vibration wheel model is established, and the vibration wheel is set as a rigid body; infinite element cells are set around and at the bottom of the model to prevent the reflection of vibration waves at the boundary from affecting the calculation results.
3. The intelligent control method for the rolling process of dam filling material according to claim 1, characterized in that The indoor tests include static and dynamic triaxial tests, shear tests, and vibration compaction tests of the fill materials, which are used to obtain various physical and mechanical parameters required for the fill materials in the numerical simulation model.
4. The intelligent control method for the rolling process of dam filling material according to claim 1, wherein For the on-site compaction tests, multiple groups of on-site compaction tests are carried out for typical working conditions. During the compaction process, the acceleration of the vibration wheel and the interaction force between the vibration wheel and the fill are measured, and the fill density and settlement deformation are measured.
5. The intelligent control method for the rolling process of dam filling material according to claim 1, characterized in that, Calibrating the parameters of the numerical simulation model specifically includes: Adjusting the parameters of the numerical simulation model according to the on-site test results so that the error between the simulation calculation results of the numerical simulation model and the on-site measured results is within 10%.
6. The intelligent control method for the rolling process of dam filling material according to claim 1, characterized in that The specific implementation of the multi-condition fill compaction simulation calculation is as follows: Set different levels for the compaction parameters and the fill compaction state, adopt orthogonal sampling, uniform sampling, random sampling, and Latin hypercube sampling to design a multi-condition test plan and conduct simulation calculations, and extract the calculation results of the vibration wheel acceleration, the interaction force between the vibration wheel and the fill, and the compaction quality under different compaction conditions.
7. An intelligent control method for the rolling process of dam filling materials according to claim 1, characterized in that, For the compaction quality evaluation surrogate model, the model inputs are the vibration wheel acceleration eigenvalue, the interaction force between the vibration wheel and the fill, the compaction vehicle speed, the vibration frequency, and the excitation force, and the model output is the current fill compaction quality. The model is as follows: ; Wherein: is the compaction quality of the filler after the i-th pass of rolling; A i is the eigenvalue of the vibratory drum acceleration during the i-th pass of rolling; F si is the interaction force between the vibratory drum and the filler during the i-th pass of rolling; v i is the rolling vehicle speed during the i-th pass of rolling; f i is the vibration frequency during the i-th pass of rolling; F ei is the exciting force during the i-th pass of rolling; For the compaction quality increment prediction proxy model, the input is the compaction quality of the filler after the current i-th pass of rolling, the rolling vehicle speed, vibration frequency, and excitation force used in the next (i + 1)-th pass of rolling, and the output is the compaction quality increment of the filler in the (i + 1)-th pass of rolling. The model is as follows: ; Wherein: is the increment of the compaction quality of the filler after the (i + 1)-th pass of rolling; is the rolling vehicle speed of the (i + 1)-th pass of rolling; is the vibration frequency of the (i + 1)-th pass of rolling; is the exciting force of the (i + 1)-th pass of rolling.
8. The intelligent control method for the rolling process of dam filling material according to claim 1, wherein, The compaction quality assessment system is specifically used for: During the rolling process, the signal acquisition system continuously acquires the original acquisition data of the dynamic response of the vibratory roller - filler, and performs preprocessing to obtain the preprocessed data, including the vibration wheel acceleration eigenvalue, the interaction force between the vibration wheel and the filler, the rolling vehicle speed, the vibration frequency, and the excitation force; inputs the preprocessed data into the compaction quality assessment proxy model to output the current compaction quality of the filler.
9. The intelligent control method for the rolling process of dam filling material according to claim 1, characterized in that, Step 8 is specifically: The rolling parameter optimization system aims to maximize the efficiency of the compaction quality increment per unit of filler, that is: ; Wherein: is the compaction efficiency of the (i + 1)-th pass of rolling; is the increment of the compaction quality of the filler after the (i + 1)-th pass of rolling, which is calculated by the compaction quality increment prediction proxy model; is the rolling vehicle speed of the (i + 1)-th pass of rolling; The constraint conditions include: (1) The constraint of not occurring chaotic vibration; The eigenvalue A of the vibratory wheel acceleration for the i-th pass of rolling i is subjected to fast Fourier transform and then spectral analysis to extract the amplitudes A 1 / 2 , A 3 / 2 , A 5 / 2 , A 7 / 2 ··· of the half harmonics 1 / 2f0, 3 / 2f0, 5 / 2f0, 7 / 2f0 ···, where f0 is the frequency; the amplitudes of the half harmonics are accumulated, and the accumulated value is not greater than the set value , that is, the following formula is satisfied: ; (2) The vibration frequency, rolling vehicle speed, and excitation force are within the adjustment range of the vibratory roller, that is, they satisfy the following formula: ; Where: f min , v min and F emin are respectively the minimum vibration frequency, the minimum rolling speed and the minimum exciting force of the roller; f max , v max and F emax are respectively the maximum vibration frequency, the maximum rolling speed and the maximum exciting force of the roller; (3) The constraint of meeting the compaction uniformity requirement: The horizontal displacement of the roller within one vibration period, that is, the impact spacing, needs to meet the following requirements: ; Wherein: is the rolling vehicle speed for the (i + 1)-th pass of rolling; is the vibration frequency for the (i + 1)-th pass of rolling; is the minimum value of the impact spacing; is the maximum value of the impact spacing; The rolling parameter optimization is implemented using the particle swarm algorithm, including: S1, Initialize the particle swarm parameters, including the particle swarm size, particle dimension, number of iterations, inertia weight, learning factor, and iteration step size; S2, Randomly initialize the position and velocity of each particle; S3, Calculate the fitness of each particle using the following formula; ; Wherein: is the fitness; is the increment of the compaction quality of the filler after the next pass, i.e., the (i + 1)-th pass, which is calculated by the compaction quality increment prediction proxy model; is the rolling vehicle speed of the next pass, i.e., the (i + 1)-th pass; S4, Update the historical optimal fitness and position of each particle and the population; S5, Update the particle velocity and position; ; Wherein: is the velocity of the i-th particle at the (k + 1)-th iteration; is the weight coefficient; is the velocity of the i-th particle at the k-th iteration; 、 are learning factors; 、 are random numbers taking values from two normal distributions, with a range of [0, 1]; is the solution corresponding to the optimal fitness obtained by the i-th particle up to the k-th iteration; is the solution corresponding to the optimal fitness obtained among all particles up to the k-th iteration; 、 are the solutions corresponding to all particles at the k-th and (k + 1)-th iterations respectively; When the new position of the particle meets the constraint conditions, update the individual position; when the new position of the particle does not meet the constraint conditions, do not update the individual position; S6, Determine whether the end condition is met; S7, If it is met, output the optimal solution; if it is not met, return to step S3 to repeat the process.
10. An intelligent control system for the rolling process of dam filling materials, characterized in that, It includes a signal acquisition system, a compaction quality assessment system, a rolling parameter optimization system, and an electronic control system; The signal acquisition system, through the satellite customization device, acceleration sensing device, and vibration wheel - filler interaction detection device installed on the vibratory roller, continuously acquires the roller position information, vibration wheel acceleration signal, and vibration wheel - filler interaction information during the vibratory rolling process, and inputs them into the compaction quality assessment system; The compaction quality assessment system preprocesses the data acquired by the signal acquisition system to obtain the vibration wheel acceleration eigenvalue, the interaction force between the vibration wheel and the filler, the rolling vehicle speed, the vibration frequency, and the excitation force, and then inputs them into the compaction quality assessment proxy model to output the current compaction quality of the filler; inputs the current compaction quality of the filler into the rolling parameter optimization system; The rolling parameter optimization system, based on the current compaction quality of the filler, aims to maximize the compaction efficiency in the next pass of rolling, and uses the particle swarm algorithm to optimize the rolling parameters in the next pass. Among them, the compaction quality increment prediction proxy model is used to calculate the particle fitness, and the optimal rolling parameters calculated by the rolling parameter optimization system are input into the electronic control system; The electric control system automatically adjusts the mechanical and hydraulic systems of the roller according to the obtained optimal compaction parameters, and controls the compaction vehicle speed, vibration frequency, and excitation force to the optimal compaction parameters.
Citation Information
Patent Citations
A numerical simulation method of a vehicle load based on a finite element-infinite element model
CN109033566A
Intelligent compaction decision system for road beds and road surfaces
CN109468918A
Comprehensive test system and method for high-speed railway roadbed
CN111562077A
Whole-process intelligent compaction system based on digital twinning technology and management and control method
CN116430739A
Road nondestructive testing method based on absorption boundary
CN116626163A
Cited By
Roadbed filling quality parameter optimization analysis system
CN120724096A
