Efficient and environmentally friendly sand screening device control system

Abstract

The present application belongs to the technical field of equipment control, and particularly relates to a high-efficiency environment-friendly sand and gravel screening equipment control system, to solve the technical problem that the existing method lacks deep understanding and accurate prediction of the internal mechanism of the screening process, and cannot realize the collaborative optimization of multiple key performance indicators such as screening efficiency, hole blocking risk and operation energy consumption. It comprises: a calculation module, which predicts the evolution trend of the hole blocking probability of the screen surface and calculates the transition energy consumption of the particles; a construction module, which constructs a weighted multi-objective optimization function; and a regulation and control module, which is used for real-time regulation and control of the screen parameters of the screening equipment according to the obtained target screen parameter combination. The present application ensures that the particle size distribution of the environment-friendly sand and gravel is highly stable and continuously meets the engineering standards, and improves the product qualification rate.

Description

Technical Field

[0001] This invention belongs to the technical field of equipment control, specifically relating to a control system for a high-efficiency and environmentally friendly sand and gravel screening equipment. Background Technology

[0002] Aggregates are indispensable basic building materials in fields such as concrete, asphalt mixtures, and roadbed engineering. Their particle size distribution, cleanliness, and other quality indicators are related to the performance and durability of the project. In the production and processing of aggregates, screening is the core step in achieving the classification and separation of particulate materials according to particle size.

[0003] Currently, the operating parameters of vibrating screen screening equipment are largely set based on manual experience, and are typically adjusted once during the equipment commissioning phase and then remain unchanged for a long period. However, in actual production, the key characteristics of the raw materials to be screened, such as particle size distribution, moisture content, mud content, and feed rate, are constantly changing due to factors such as the source, mining batch, and weather. This traditional control method leads to the equipment's inability to adapt to fluctuations in material characteristics in real time, often resulting in problems such as low screening efficiency, unstable finished product quality, and deviations from the target gradation range.

[0004] Especially when the material has high moisture content or a large number of critical-sized particles, screen clogging is likely to occur. This not only reduces the screening area and processing capacity but also requires frequent shutdowns for manual cleaning, severely impacting production continuity and economic efficiency. Furthermore, blindly increasing vibration intensity to alleviate clogging results in significant energy waste. While existing methods can adjust the feed rate by monitoring the outlet material flow or determine the load by monitoring the motor current, these methods mostly involve indirect and delayed adjustments to single parameters, lacking an understanding and accurate prediction of the internal mechanisms of the screening process.

[0005] The screening process is essentially a process involving particle group movement and energy transfer. In particular, the evolution trend of the probability of screen clogging is influenced by the combined effects of material characteristics, screen parameter combinations, and particle transition behavior on the screen surface, making it difficult to represent with a simple linear model or empirical relationship. Therefore, traditional control strategies generally cannot achieve synergistic optimization of multiple key performance indicators such as screening efficiency, clogging risk, and operating energy consumption.

[0006] Furthermore, for new equipment such as cross screens, multiple sets of vibrators work together to form a special interference wave field to enhance the screening effect. This places higher demands on the precise and coordinated control of the phase spectrum of multiple vibration sources, which traditional control methods are simply not up to the task.

[0007] Therefore, developing an advanced control method that can sense changes in operating conditions in real time, predict process trends, and make multi-objective intelligent optimization decisions has become an urgent need to promote the development of the sand and gravel screening industry towards high efficiency, environmental protection, and intelligence. Summary of the Invention

[0008] This invention provides a high-efficiency and environmentally friendly sand and gravel screening equipment control system to solve the technical problem that existing methods lack a deep understanding and accurate prediction of the internal mechanism of the screening process, and cannot achieve synergistic optimization of multiple key performance indicators such as screening efficiency, clogging risk and operating energy consumption.

[0009] A high-efficiency and environmentally friendly sand and gravel screening equipment control system includes:

[0010] The calculation module is used to obtain the particle size distribution, humidity, feeding speed of the inlet material and the particle size distribution of each outlet material, and to calculate the screening efficiency; using the screening process prediction model, based on real-time data and the target screen parameters of the previous cycle, it predicts the evolution trend of the probability of screen blockage and calculates the transition energy consumption of particles.

[0011] The module is used to construct a weighted multi-objective optimization function with the objectives of maximizing screening efficiency, minimizing the fluctuation of clogging probability, and minimizing transition energy consumption. The weight coefficients are adjusted according to the deviation between the target screening parameters of the previous cycle and the preset benchmark parameters. In each control cycle, the weighted multi-objective optimization function is solved by a multi-island genetic algorithm to obtain the target screening parameter combination for the current control cycle, and the target screening parameter combination of the previous cycle is retained as an elite individual.

[0012] The solution process satisfies the following conditions: when the screening equipment is a cross screen, the cooperative vibration phase spectrum required to form the interference wave field is a constraint condition; when the second derivative of the particle size distribution of the finished product exceeds the preset instability threshold, the population diversity is enhanced by increasing the mutation probability or inputting random individuals.

[0013] The control module is used to adjust the screening parameters of the screening equipment in real time based on the obtained target screening parameter combination.

[0014] Furthermore, the real-time particle size distribution, humidity, and feed rate of the inlet material, as well as the particle size distribution of each outlet material, are obtained, and the screening efficiency is calculated, including:

[0015] An infrared moisture meter is installed at the feed inlet of the screening equipment to continuously measure the moisture content of the incoming material;

[0016] A weighing sensor is installed below the feed conveyor belt, and the feeding speed is calculated in combination with the belt speed.

[0017] Laser particle size analyzers are installed at the feed inlet and each finished product outlet to acquire real-time particle size distribution data of the material before and after it passes through.

[0018] The screening efficiency is calculated based on the percentage of particles smaller than the screen aperture in the material fed into the screen and the product on the screen.

[0019] Furthermore, the screening efficiency satisfies the following relationship:

[0020] ;

[0021] In the formula, For screening efficiency, This represents the percentage of particles smaller than the sieve aperture in the material being screened. This represents the percentage of particles smaller than the sieve aperture in the product passed through the sieve.

[0022] Furthermore, using a screening process prediction model, based on real-time data and the target screen parameter combination from the previous cycle, the evolution trend of the screen surface clogging probability is predicted, and the particle transition energy consumption is calculated, including:

[0023] The screen surface is divided into multiple grid cells. The motion trajectory and collision events of each particle are tracked based on the discrete element method. When the dwell time of a particle in a single grid cell exceeds a preset dwell time threshold, the clogging probability of the corresponding grid cell is increased by a fixed increment. The transition energy consumption of the particle is calculated and accumulated based on the changes in kinetic and potential energy when the particle successfully transitions from one layer of screen to the next.

[0024] Furthermore, the energy consumption of particle transitions satisfies the following relationship:

[0025] ;

[0026] In the formula, It is the energy consumption of particle transition; It is the change in kinetic energy; It is the change in potential energy; It refers to particle mass. and These are the instantaneous velocities of the particles before and after passing through the sieve holes. and These are the vertical positions of the particles before and after passing through the sieve holes.

[0027] Furthermore, the weighting coefficients are adjusted based on the deviation between the target screening parameters and the preset benchmark parameters from the previous period, including:

[0028] Calculate the Euclidean distance between the target sieve parameter vector and the preset reference parameter vector in the previous cycle; use this Euclidean distance as a variable to linearly and positively adjust the weights of the clogging probability fluctuation and transition energy consumption, and linearly and negatively adjust the weights of the screening efficiency, so that when the target sieve parameter vector in the previous cycle deviates from the preset reference parameter vector, the weights of the clogging probability fluctuation and transition energy consumption are increased.

[0029] Furthermore, in each control cycle, a multi-island genetic algorithm is used to solve the weighted multi-objective optimization function to obtain the target sieve parameter combination for the current control cycle, while the target sieve parameter combination from the previous cycle is retained as elite individuals, including:

[0030] The entire population is divided into several independent islands, and each island independently performs selection, crossover, and mutation operations. At the beginning of each control period, the target screening parameter combination calculated in the previous period is input into the population of the current period as elite individuals, and parallel evolutionary computation of all islands is initiated.

[0031] Furthermore, the solution process satisfies the following constraint condition when the screening equipment is a cross-screen: the cooperative vibration phase spectrum required to form the interference wave field includes:

[0032] For the two exciters of the cross screen, the vibration phase difference satisfies the relationship that is proportional to the vibration frequency and the center distance between the two exciters, and inversely proportional to the propagation speed of the vibration wave on the screen surface; in the fitness function evaluation stage of the multi-island genetic algorithm, for individuals that do not meet the phase difference constraint, a penalty term is applied to the fitness value, and the value of the penalty term is proportional to the absolute difference between the phase difference of the individual solution and the required phase difference.

[0033] Furthermore, when the second derivative of the particle size distribution of the exported finished material exceeds a preset instability threshold, population diversity is enhanced by increasing the mutation probability or inputting random individuals, including:

[0034] The particle size distribution curve of the exported finished material is monitored in real time, and the maximum value of the second derivative is calculated. If the maximum value exceeds the preset instability threshold, the mutation probability is increased from a lower base value to a higher value in the subsequent multi-island genetic algorithm iterations, and several completely randomly generated new individuals are forcibly input into each island in each generation to escape the local optimum.

[0035] Furthermore, based on the obtained target sieve parameter combination, the sieve parameters of the screening equipment are adjusted in real time, including:

[0036] The optimized target screen parameter combination is decomposed into specific control commands. The target screen parameter combination includes vibration frequency, amplitude, screen surface inclination angle, and vibration phase difference of cross screen.

[0037] The programmable logic controller outputs a PWM signal to the frequency converter to adjust the speed of each vibrator motor to control the vibration frequency and vibration phase difference; the amplitude is adjusted by adjusting the relative position of the eccentric blocks; and the servo motor is driven to adjust the support angle of the screen box through the linkage mechanism to achieve real-time control of the screen surface inclination angle.

[0038] The beneficial effects are as follows: This invention achieves simultaneous improvement in screening efficiency, finished product quality stability, and operating energy consumption through precise prediction and global optimization of the internal mechanism of the screening process. By predicting the evolution trend of the probability of screen blockage and calculating the transition energy consumption of particles, the system can adjust operating parameters before blockage occurs to avoid it, reducing downtime for cleaning due to screen blockage and ensuring production continuity. By synergistically optimizing multiple mutually restrictive key indicators such as maximizing screening efficiency, minimizing the fluctuation of blockage probability, and minimizing transition energy consumption, it can find a balance point with optimal comprehensive performance under the current material characteristics, ensuring that the particle size distribution of environmentally friendly sand and gravel is highly stable and continuously meets engineering standards, thus improving the product qualification rate. Especially for complex equipment such as cross screens, it ensures that multiple sets of vibrators work together in the optimal phase relationship, thereby leveraging the technical advantages of enhanced screening. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of a LSTM-based prediction model for the screening process.

[0040] Figure 2 A schematic diagram of the structure of a weighted multi-objective optimization function;

[0041] Figure 3 This is a diagram showing the connection relationship between the PLC controller and the actuator. Detailed Implementation

[0042] An embodiment of the high-efficiency and environmentally friendly sand and gravel screening equipment control system provided by the present invention:

[0043] A high-efficiency and environmentally friendly sand and gravel screening equipment control system includes a calculation module, a construction module, and a control module.

[0044] The calculation module is used to obtain the real-time particle size distribution, humidity, and feeding speed of the inlet material and the real-time particle size distribution of each outlet material, and to calculate the screening efficiency. Using the screening process prediction model, based on the real-time data and the target screen parameters of the previous cycle, it predicts the evolution trend of the probability of screen blockage and calculates the transition energy consumption of particles.

[0045] Specifically, an online particle size analyzer based on machine vision is installed above the feed belt of the screening equipment. The camera continuously captures images of the material flow, and image processing algorithms are used to identify particle outlines and statistically analyze the particle size distribution of the inlet and outlet materials. A microwave humidity sensor installed above the feed belt measures the material's moisture content in real time. A belt scale monitors the feeding speed in real time. Similar online particle size analyzers are also installed above the outlet belts of each finished product material to obtain the particle size distribution of the outlet material. The particle size distribution of the inlet material, moisture content, feeding speed, and particle size distribution of the outlet material together constitute the real-time data.

[0046] Based on the law of conservation of mass, the controller uses the feed rate and particle size distribution data of each outlet material, and calculates the screening efficiency in real time according to the general formula for screening efficiency, which is the ratio of the target particle size content in the undersize product to the target particle size content in the feed.

[0047] In an optional embodiment, the screening process prediction model is a Long Short-Term Memory (LSTM) network pre-trained with a large amount of experimental data and discrete element simulation data. The controller uses parameters such as vibration frequency, amplitude, and screen tilt angle set in the previous control cycle, along with real-time data collected in the current control cycle, as the input vector to the screening process prediction model. The model outputs a curve showing the probability of clogging caused by critical-sized particles on the screen over a future time period, i.e., the evolution trend of the clogging probability. An embedded energy consumption calculation sub-model based on particle dynamics calculates the theoretical energy required for a single particle to complete one projectile motion on the screen surface, based on the input screen parameters and material characteristics, and estimates the overall particle transition energy consumption by combining it with the feeding rate. Figure 1 As shown.

[0048] In an optional embodiment, the real-time particle size distribution, humidity, feeding rate of the inlet material, and particle size distribution of each outlet material are obtained, and the screening efficiency is calculated, including:

[0049] An infrared moisture meter is installed at the feed inlet of the screening equipment to continuously measure the moisture content of the incoming material;

[0050] A weighing sensor is installed below the feed conveyor belt, and the feeding speed is calculated in combination with the belt speed.

[0051] Laser particle size analyzers are installed at the feed inlet and each finished product outlet to acquire real-time particle size distribution data of the material before and after it passes through.

[0052] The screening efficiency is calculated based on the percentage of particles smaller than the screen aperture in the material fed into the screen and the product on the screen.

[0053] Specifically, in one example, an infrared moisture meter is installed above the feeding conveyor belt at the inlet of the screening equipment. This instrument continuously emits infrared light of a specific wavelength and detects the reflected light, thereby calculating the moisture content of the material in real time. For example, the moisture content of the material is 5.8%. A load cell installed below the conveyor belt measures the weight of the material per unit length in real time, and combined with the belt speed measured by the encoder, the feeding speed is calculated. For example, the belt speed is 1.5. The feeding speed is 100 .

[0054] Laser particle size analyzers are deployed at both the feed inlet and the two discharge outlets for oversize and undersize materials. Through laser diffraction or dynamic image analysis technology, complete particle size distribution data is updated once per second. For example, the median particle size of the inlet material is 15 mm, the median particle size of the oversize material is 25 mm, and the median particle size of the undersize material is 8 mm.

[0055] Screening efficiency is a core indicator for evaluating the effectiveness of the screening process, and its calculation depends on the mass balance of the critical particle size distribution. According to international standards, screening efficiency is typically calculated based on the content of particles (i.e., particles smaller than the sieve aperture) in both the feed material and the oversize product. Screening efficiency satisfies the following relationship: In the formula, For screening efficiency, This represents the percentage of particles smaller than the sieve aperture in the material being screened. This represents the percentage of particles smaller than the sieve aperture in the product passed through the sieve.

[0056] For example, if the sieve aperture is 10mm, and a laser particle size analyzer measures that 40% of the material entering the sieve contains particles smaller than 10mm, and 5% of the product remaining on the sieve contains particles smaller than 10mm, then the sieving efficiency is: .

[0057] In an optional embodiment, a screening process prediction model is used to predict the evolution trend of the screen surface clogging probability based on real-time data and the target screen parameter combination of the previous cycle, and to calculate the particle transition energy consumption, including:

[0058] The screen surface is divided into multiple grid cells. The motion trajectory and collision events of each particle are tracked based on the discrete element method. When the dwell time of a particle in a single grid cell exceeds a preset dwell time threshold, the clogging probability of the corresponding grid cell is increased by a fixed increment. The transition energy consumption of the particle is calculated and accumulated based on the changes in kinetic and potential energy when the particle successfully transitions from one layer of screen to the next.

[0059] Specifically, the model structure for predicting the evolution trend of screen clogging probability is a simulation model based on the Discrete Element Method (DEM). The screen surface is divided into a 100×50 grid cell in virtual space. The DEM treats each particle as an independent computational unit, tracking its trajectory on the screen surface by solving Newton's second law, and handling collisions between particles and between particles and the screen surface. To achieve this simulation process, two core prerequisites must be clarified: first, how to obtain the real-time motion parameters of the microscopic particles of sand and gravel to provide accurate input for the simulation; and second, how to determine the model input parameters that dynamically change with material properties to ensure that the simulation matches actual working conditions.

[0060] For real-time measurement of the motion parameters of microscopic particles in sand and gravel, this embodiment adopts a multi-sensor fusion scheme: two sets of high-speed industrial cameras are symmetrically installed on both sides of the screen box, and a laser supplementary lighting system is used to cover the entire screen surface, ensuring complete capture of the motion state of sand and gravel particles on the screen surface; three sets of laser Doppler velocimeters are evenly arranged below the screen along the material movement direction, corresponding to the areas through which sand and gravel particles pass at the feed end, middle, and discharge end of the screen, respectively, to monitor the movement speed of sand and gravel in the initial, middle, and final stages of screening, and comprehensively capture the dynamic changes in the motion state of sand and gravel.

[0061] The images of sand and gravel particles acquired by a high-speed industrial camera are processed through the following steps: 1) Grayscale processing, such as using a weighted average method to enhance the contrast between the sand and gravel particles and the background; 2) Threshold segmentation, such as using the maximum inter-class variance method to automatically determine the threshold and separate the sand and gravel particles from the screen surface and impurities; 3) Morphological processing, such as using opening operations to remove noise and closing operations to fill holes, to optimize the contours of the sand and gravel particles; 4) Extracting the pixel coordinates of the sand and gravel particles through a particle tracking velocimetry algorithm, and converting them into physical coordinates by combining the camera calibration parameters to obtain the motion trajectory of a single sand and gravel particle.

[0062] The particle velocity data of sand and gravel collected by the laser Doppler velocimeter are obtained using 3D... The criteria include removing outliers and eliminating abnormal data caused by sensor interference or particle collisions; weighted fusion of the sand and gravel particle velocity data extracted from the images with the velocity data measured by laser to obtain the instantaneous velocity of the sand and gravel particles before and after passing through the screen holes; and analysis of the vertical position changes of the sand and gravel particles through the image sequence acquired by the camera, combined with the screen installation height, to determine the vertical position of the sand and gravel particles before and after passing through the screen holes.

[0063] Regarding the dynamic determination of model input parameters, considering the continuous fluctuations in material properties during actual production, this embodiment establishes an adaptive adjustment mechanism: the dynamic input parameters of the model include two categories: first, material characteristic parameters, such as particle size distribution standard deviation, humidity, and mud content; second, equipment operating status parameters, such as feeding speed and screen surface vibration acceleration. All parameters are sampled at a frequency of 10Hz to ensure real-time performance. The mud content is estimated using a preset humidity-mud content mapping table, which is based on more than 500 sets of experimental data and can quickly respond to differences in material properties caused by changes in humidity.

[0064] To achieve accurate matching between input parameters and discrete element model coefficients, a BP neural network model was constructed: the input layer has 4 neurons (particle size distribution standard deviation, moisture content, mud content, and feed rate), the hidden layer has 2 layers, and the output layer has 3 neurons (contact stiffness, coefficient of restitution, and coefficient of friction). The neural network was trained using experimental data from over 500 sets of different material conditions. In one example, the training parameters were: learning rate 0.01, number of iterations 1000, and loss function: mean squared error. Boundary constraints were also set for the parameters, such as moisture content. mud content Feeding speed When the sensor data exceeds this range, the moving average of the first 5 valid data sets is used as a temporary input parameter, and an alarm signal is issued by the PLC to prompt for inspection. Every 20 new valid data sets are collected, the BP neural network is incrementally trained to ensure that the mapping relationship dynamically matches the actual material characteristics.

[0065] Based on the measured particle motion parameters and dynamically determined model input parameters, discrete element simulation can be successfully carried out. In this embodiment, a soft ball contact model is used. The friction coefficient between particles is dynamically adjusted according to the material type. For example, it is 0.4 for ordinary sand and gravel, and 0.5 when the mud content is greater than 10%. The friction coefficient between the particles and the screen surface is fixed at 0.3. The time step is 1 / 4 of the critical time step, which is determined by the formula... Calculate, where, For the minimum particle mass, for example =0.001kg, For maximum contact stiffness, for example =5× To ensure a balance between simulation stability and real-time performance, a single control cycle is set to 50ms, within which particle motion tracking, collision event handling, and hole blockage probability calculation for a 100×50 mesh element are completed.

[0066] In predicting the probability of clogging, a basic residence time threshold is set, for example, 5 seconds. Simultaneously, the threshold is dynamically adjusted by comprehensively considering the matching degree between particle size and sieve aperture size, as well as material moisture content. The matching degree is the ratio of particle size to sieve aperture size. When the matching degree... Furthermore, when the material moisture content is greater than 10%, the residence time threshold is adjusted to 3 seconds. The clogging probability increment is dynamically set based on the matching degree; for example, when the matching degree is 1.0, the clogging probability increment is 0.02, and when the matching degree is 0.8 or 1.2, the clogging probability increment is 0.005. One pressure sensor is installed for each grid unit on the screen surface to monitor the pressure value of the grid unit in real time. When the pressure value exceeds a preset threshold, the grid unit is determined to be clogged. Every two control cycles, the error between the predicted clogging probability and the actual clogging state is calculated. If the error is greater than 5%, the next predicted clogging probability increment is corrected using a proportional coefficient, for example, a proportional coefficient of 0.1.

[0067] When the probability of clogging in a certain grid cell is predicted to exceed 80% within the next 3 control cycles, parameter adjustment is immediately triggered: the vibration frequency is adaptively increased by 0.5-1Hz based on the current value, and the amplitude is increased by 0.2-0.5mm accordingly. During the adjustment process, the screening efficiency is ensured not to be lower than the benchmark value. If the clogging efficiency continues to increase for 2 consecutive control cycles after adjustment, the purging device needs to be started to assist in cleaning the holes.

[0068] Calculating particle transition energy consumption is one of the model's functions, used to represent the energy efficiency of the sieving process. When a particle successfully passes through the sieve openings from one side to the other, its mechanical energy changes; this change is defined as the particle's transition energy consumption, which satisfies the following relationship: In the formula, It is the energy consumption of particle transition; It is the change in kinetic energy; It is the change in potential energy; It refers to particle mass. and These are the instantaneous velocities of the particles before and after passing through the sieve holes. and These are the vertical positions of the particles before and after passing through the sieve holes.

[0069] For example, consider a particle with a mass of 0.002 kg. Its instantaneous velocity before piercing is 0.5 m / s, and its instantaneous velocity after piercing is 1.2 m / s, with a height decrease of 0.01 m. The energy consumption can then be calculated. The model accumulates the particle transition energy consumption of all particles that successfully pass through the screen throughout the simulation period to obtain the total transition energy consumption. The total transition energy consumption is one of the objectives to be minimized in multi-objective optimization, and it satisfies the following relationship: In the formula, It is the total transition energy consumption. It represents the total number of particles that successfully passed through the sieve during the simulation period.

[0070] The construction module is used to construct a weighted multi-objective optimization function with the objectives of maximizing screening efficiency, minimizing the fluctuation of clogging probability, and minimizing transition energy consumption. The weight coefficients are adjusted according to the deviation between the target screening parameters of the previous cycle and the preset benchmark parameters. In each control cycle, the weighted multi-objective optimization function is solved by a multi-island genetic algorithm to obtain the target screening parameter combination for the current control cycle, and the target screening parameter combination of the previous cycle is retained as an elite individual.

[0071] The solution process satisfies the following conditions: when the screening equipment is a cross screen, the cooperative vibration phase spectrum required to form the interference wave field is a constraint condition; when the second derivative of the particle size distribution of the finished product exceeds the preset instability threshold, the population diversity is enhanced by increasing the mutation probability or inputting random individuals.

[0072] Specifically, the controller stores a set of preset benchmark sieve parameters representing the highest energy efficiency ratio. Before the optimization begins in each cycle, the Euclidean distance between the target sieve parameters used in the previous cycle and the preset benchmark parameters is calculated. The larger the Euclidean distance, the more likely the screening equipment is operating in an uneconomical zone. At this time, the weighting coefficient is adjusted.

[0073] In the fitness evaluation stage of the genetic algorithm, for cross-screening, each individual represents a combination of screen parameters containing the vibration phases of each exciter. The program checks whether this set of vibration phases satisfies the physical constraint equations for forming a specific interference wave field, such as linear or elliptical vibration. If not, the individual is assigned a very low fitness value, causing it to be eliminated in subsequent selections. Simultaneously, the controller continuously monitors the particle size distribution curve of the finished material and calculates the second derivative. If the peak value of the second derivative exceeds a preset threshold, it indicates an abnormal spike in the particle size distribution, potentially indicating a local optimum. The algorithm will automatically increase the mutation probability from 1% to 5%, or randomly generate a new batch of individuals from the population to replace those with poor fitness, thereby escaping the local optimum and re-exploring a better operating region.

[0074] In an optional embodiment, the weighting coefficient is adjusted based on the deviation between the target screening parameter and the preset benchmark parameter in the previous period, including:

[0075] Calculate the Euclidean distance between the target sieve parameter vector and the preset reference parameter vector in the previous cycle; use this Euclidean distance as a variable to linearly and positively adjust the weights of the clogging probability fluctuation and transition energy consumption, and linearly and negatively adjust the weights of the screening efficiency, so that when the target sieve parameter vector in the previous cycle deviates from the preset reference parameter vector, the weights of the clogging probability fluctuation and transition energy consumption are increased.

[0076] Specifically, the construction of the weighted multi-objective optimization function aims to balance three mutually constraining performance metrics, which satisfy the following relationship: In the formula, The negative value represents the screening efficiency, i.e., 1- ; The standard deviation represents the probability of clogging of all grid cells on the screen surface. The standard deviation indicates that the more uniform the clogging distribution, the more stable the working state of the screen surface. , , These are the weight coefficients of the three objectives, and their sum is 1.

[0077] The dynamic adjustment mechanism of the weighting coefficients is designed to achieve control by setting an ideal preset combination of parameters: the preset parameter vector is a frequency of 15Hz, an amplitude of 5mm, and a screen inclination angle of 18°. At the end of each control cycle, the target screen parameter vector currently in use is recorded; for example, a frequency of 18Hz, an amplitude of 6mm, and a screen inclination angle of 20°. The Euclidean distance between the two vectors is calculated, and the weights for the next cycle are linearly adjusted based on this distance. For example, the weights for screening efficiency, clogging probability fluctuation, and transition energy consumption are updated according to the following relationship: ; ; In the formula, As the efficiency weight for the next cycle, For the stable weights of the next cycle, As the energy consumption weight for the next cycle, This is the baseline weight value for screening efficiency under ideal operating conditions. This serves as the baseline weighting value for the fluctuation in plugging efficiency under ideal operating conditions. This is the baseline weight value for transition energy consumption under ideal operating conditions. For positive integers, For Euclidean distance. These relationships illustrate that the further the actual operating parameters deviate from the ideal baseline, the greater the focus will be on operational stability and energy consumption, automatically reducing the pursuit of ultimate screening efficiency, such as... Figure 2 As shown.

[0078] In an optional embodiment, in each control cycle, a multi-island genetic algorithm is used to solve a weighted multi-objective optimization function to obtain the target sieve parameter combination for the current control cycle, and the target sieve parameter combination from the previous cycle is retained as elite individuals, including:

[0079] The entire population is divided into several independent islands, and each island independently performs selection, crossover, and mutation operations. At the beginning of each control period, the target screening parameter combination calculated in the previous period is input into the population of the current period as elite individuals, and parallel evolutionary computation of all islands is initiated.

[0080] Specifically, the multi-island genetic algorithm model structure promotes diversity by dividing the total candidate solution population into four independent islands, each containing 50 individuals. Each individual is encoded with a set of screening parameters, such as a frequency of 15 Hz, an amplitude of 5 mm, and an inclination angle of 18°. Within each island, independent genetic algorithm operations, including single-point crossover and basic bit mutation, are performed to simulate independent evolutionary processes. This parallel structure allows different subpopulations to explore different regions of the search space, preventing the algorithm from prematurely getting trapped in local optima.

[0081] In each new control cycle, an elite retention strategy is implemented. The optimal sieve parameter combination found at the end of the previous control cycle, such as frequency 16.2Hz, amplitude 4.8mm, and tilt angle 17.5°, is copied as an elite individual to the initial population of the current cycle, typically replacing one of the worst individuals. This ensures that the algorithm's evolutionary process does not lose the discovered optimal solutions. After several generations of parallel evolution across all islands, a migration event occurs. Each island selects several of the best individuals and sends their copies to neighboring islands, replacing inferior individuals in neighboring islands. This achieves the exchange of superior genes between different subpopulations, accelerating the discovery of the global optimum.

[0082] In an optional embodiment, the solution process satisfies the following constraint condition when the screening equipment is a cross-screen: the cooperative vibration phase spectrum required to form the interference wave field includes:

[0083] For the two exciters of the cross screen, the vibration phase difference satisfies the relationship that is proportional to the vibration frequency and the center distance between the two exciters, and inversely proportional to the propagation speed of the vibration wave on the screen surface; in the fitness function evaluation stage of the multi-island genetic algorithm, for individuals that do not meet the phase difference constraint, a penalty term is applied to the fitness value, and the value of the penalty term is proportional to the absolute difference between the phase difference of the individual solution and the required phase difference.

[0084] Specifically, for cross screens using two or more exciters, in order to form a specific interference wave field on the screen surface and improve screening efficiency, the vibration phase difference between the two exciters must satisfy a specific condition, which satisfies the following physical formula: In the formula, For the vibration phase difference, It is the vibration frequency. It is the straight-line distance between the centers of the two exciters. It is the propagation speed of the vibration wave on the screen surface medium.

[0085] For example, in a cross-screen, the center distance between two vibrators is 2m, and the vibration wave propagates at a speed of 30m / s on the screen surface. If a genetic algorithm generates an individual representing a vibration frequency of 15Hz, then the required phase difference to form an ideal interference field should be 2. rad, which stands for in-phase vibration.

[0086] When evaluating the fitness of each individual, i.e., calculating the multi-objective optimization function value, the phase difference is checked to ensure it meets the constraints. The phase difference parameter is included in the individual's genetic code. The required phase difference is calculated based on the individual's frequency. If the individual's own phase difference does not match the required phase difference, a penalty is applied to the fitness value. The penalized fitness value satisfies the following formula: In the formula, The fitness value after penalty. It is the original fitness value. It is a sufficiently large penalty coefficient, for example, 1000. For the phase difference of the individual, This represents the required phase difference. If an object has a phase difference of 0.5 at a frequency of 15Hz... If the value is rad, then its fitness value will be increased by a factor of 1500. The penalty term puts individuals at a great disadvantage in the selection process, thus naturally eliminating them and ensuring that all convergent solutions satisfy the physical constraints.

[0087] In an optional embodiment, when the second derivative of the particle size distribution of the exported finished product exceeds a preset instability threshold, population diversity is enhanced by increasing the mutation probability or inputting random individuals, including:

[0088] The particle size distribution curve of the exported finished material is monitored in real time, and the maximum value of the second derivative is calculated. If the maximum value exceeds the preset instability threshold, the mutation probability is increased from a lower base value to a higher value in the subsequent multi-island genetic algorithm iterations, and several completely randomly generated new individuals are forcibly input into each island in each generation to escape the local optimum.

[0089] Specifically, the particle size distribution function of the finished product is continuously obtained using a laser particle size analyzer. ,in The particle size is The percentage of particles. To detect drastic changes or morphological anomalies in the distribution curve, the curve is numerically differentiated to obtain the second derivative. The magnitude of the peak value of the absolute value of the second derivative can indicate the curvature of the curve. An abnormally large peak indicates a drastic change in particle size distribution, suggesting that the screening process may encounter unstable states such as severe clogging. An instability threshold is preset, for example... An instability alarm is triggered if the maximum value of the second derivative calculated in real time exceeds 5.0.

[0090] Once an unstable state is detected, the control system immediately adjusts the evolutionary parameters of the genetic algorithm to cope. In subsequent algorithm iterations, the mutation probability is temporarily increased from a normally low value to a higher value. A high mutation rate increases the likelihood of generating entirely new individual genes, thereby enhancing the algorithm's exploration ability in the parameter space. Simultaneously, to further break the trap of potentially unstable local optima, a small number of completely randomly generated individuals are forced into the input for each generation of evolution on each island. For example, in an island with 50 individuals, the two worst individuals are replaced with two entirely new random individuals each generation. These measures work together to increase population diversity, helping the algorithm quickly escape currently ineffective search regions and find new parameter combinations that can restore stability to the production process.

[0091] The control module is used to adjust the screening parameters of the screening equipment in real time based on the obtained target screening parameter combination.

[0092] Specifically, the controller converts the optimal combination of screening parameters output by the multi-island genetic algorithm, such as frequency, amplitude, and tilt angle, into specific control commands. For example, the optimal frequency value is sent to the frequency converter controlling the vibration motor, the optimal amplitude value is sent to the servo motor adjusting the position of the eccentric block, and the optimal tilt angle value is sent to the controller driving the hydraulic cylinder or electric push rod to tilt the screen box, thereby completing closed-loop, real-time, and precise adjustment of the screening equipment's operating status.

[0093] In an optional embodiment, the sieve parameters of the screening equipment are adjusted in real time according to the obtained target sieve parameter combination, including:

[0094] The optimized target screen parameter combination is decomposed into specific control commands. The target screen parameter combination includes vibration frequency, amplitude, screen surface inclination angle, and vibration phase difference of cross screen.

[0095] The programmable logic controller outputs a PWM signal to the frequency converter to adjust the speed of each vibrator motor to control the vibration frequency and vibration phase difference; the amplitude is adjusted by adjusting the relative position of the eccentric blocks; and the servo motor is driven to adjust the support angle of the screen box through the linkage mechanism to achieve real-time control of the screen surface inclination angle.

[0096] Specifically, after solving the problem, the genetic algorithm outputs a set of optimal target sieve parameters, such as a target vector with a vibration frequency of 16.5 Hz, an amplitude of 5.2 mm, a sieve surface inclination angle of 18.5°, and a phase difference of 0.8. rad. This set of optimal target screening parameters is sent to the central programmable logic controller (PLC). The PLC's internal program decomposes these target screening parameters into specific set values ​​for the corresponding actuators. For example, 16.5Hz is converted into the target speed of the corresponding vibrator motor; 5.2mm is converted into the target position code for adjusting the eccentric block mechanism; 18.5° is converted into the target pulse count for adjusting the tilt servo motor; 0.8 The rad is converted to the synchronization signal delay time between the two exciter motor controllers.

[0097] For vibration frequency and phase difference, the PLC sends pulse width modulation signals to the frequency converter (VFD) controlling the two vibrator motors or communicates via fieldbus. The VFD precisely adjusts the frequency of the output power supply according to the instructions, thereby controlling the motor speed. The phase difference is precisely maintained by adjusting the phase of the slave motor relative to the master motor in real time using encoder feedback in a master-slave control mode. For amplitude, the PLC controls a small motor to drive a mechanical device to change the relative angle or position of the eccentric blocks on the vibrator, and confirms that the amplitude reaches 5.2mm through feedback from a displacement sensor. For the screen surface tilt angle, the PLC drives a high-precision servo motor, which raises or lowers one end of the screen box through a worm gear or linkage mechanism. At the same time, a tilt sensor mounted on the screen box provides real-time angle data to the PLC, forming a closed-loop control until the screen surface stabilizes at 18.5°. The entire control process is real-time, linked, and precise. Figure 3 As shown.

[0098] In addition, in the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

Claims

1. A high-efficiency environmentally friendly sand and gravel screening plant control system, characterized by, The application relates to a method for dynamically adjusting the screen parameters of a screening device, comprising the following steps: a calculation module is used to acquire the particle size distribution, humidity, feeding speed of the inlet material and the particle size distribution of each outlet material, and to calculate the screening efficiency; a screening process prediction model is used to predict the evolution trend of the screen hole blocking probability based on the real-time data and the target screen parameter combination of the last period, and to calculate the energy consumption of particle transition; a construction module is used to construct a weighted multi-objective optimization function with the maximization of screening efficiency, the minimization of hole blocking probability fluctuation and the minimization of transition energy consumption as the targets, wherein the weight coefficient is adjusted according to the deviation degree of the target screen parameter of the last period and the preset reference parameter; in each control period, the multi-island genetic algorithm is used to solve the weighted multi-objective optimization function, the target screen parameter combination of the current control period is obtained, and the target screen parameter combination of the last period is reserved as the elite individual; the solving process meets the following conditions: when the screening device is a cross screen, the cooperative vibration phase spectrum required for forming an interference wave field is a constraint condition; when the second-order derivative of the outlet product particle size distribution exceeds a preset unstable threshold, the population diversity is enhanced by increasing the mutation probability or inputting random individuals; a regulation module is used to regulate the screen parameters of the screening device in real time according to the obtained target screen parameter combination.

2. The efficient and environmentally friendly aggregate screening plant control system of claim 1, wherein, The real-time particle size distribution, humidity, feeding speed of the inlet material and the particle size distribution of each outlet material are acquired, and the screening efficiency is calculated, comprising the following steps: an infrared moisture meter is installed at the feeding inlet of the screening device to continuously measure the humidity of the inlet material; a weighing sensor is arranged below the feeding belt to calculate the feeding speed in combination with the belt speed; laser particle size analyzers are respectively installed at the feeding inlet and the outlet of each product to obtain the particle size distribution data of the materials before and after passing through in real time; the screening efficiency is calculated according to the percentage of the particles smaller than the screen hole size in the inlet material and the product on the screen.

3. The control system for the high-efficiency environmentally friendly sand and gravel screening plant of claim 2, wherein, The screening efficiency meets the following relationship: ； wherein is the screening efficiency, is the percentage of particles smaller than the screen opening size in the feed material, is the percentage of particles smaller than the screen opening size in the oversize product.

4. The control system for efficient and environmentally friendly sand and gravel screening plant according to claim 1, characterized in that, a screening process prediction model is used to predict the evolution trend of the screen hole blocking probability based on the real-time data and the target screen parameter combination of the last period, and to calculate the energy consumption of particle transition, comprising the following steps: the screen surface is divided into multiple grid units, the motion trajectory and collision events of each particle are tracked based on the discrete element method, when the residence time of the particle in a single grid unit exceeds a preset residence time threshold, the hole blocking probability of the corresponding grid unit is increased by a fixed increment; the energy consumption of particle transition is calculated and accumulated according to the change amount of kinetic energy and potential energy when the particle successfully transitions from one screen layer to the next screen layer.

5. The control system for the high-efficiency environmentally friendly sand and gravel screening plant of claim 4, wherein, The energy consumption of particle transition meets the following relationship: ； wherein is the energy consumption of the particle transition; is the change in kinetic energy; is the change in potential energy; is the mass of the particle, and are the instantaneous velocities of the particle before and after passing through the aperture, respectively, and are the vertical positions of the particle before and after passing through the aperture, respectively.

6. The efficient and environmentally friendly aggregate screening plant control system of claim 1, wherein, the weight coefficient is adjusted according to the deviation degree of the target screen parameter of the last period and the preset reference parameter, comprising the following steps: the Euclidean distance between the target screen parameter vector of the last period and the preset reference parameter vector is calculated; the weights of the hole blocking probability fluctuation and the transition energy consumption are linearly and positively adjusted, and the weight of the screening efficiency is linearly and negatively adjusted, so that when the target screen parameter vector of the last period deviates from the preset reference parameter vector, the weights of the hole blocking probability fluctuation and the transition energy consumption are increased.

7. The control system for efficient and environmentally friendly sand and gravel screening plant according to claim 1, characterized in that, In each control cycle, a multi-island genetic algorithm is used to solve a weighted multi-objective optimization function to obtain a target screen parameter combination for the current control cycle, and a target screen parameter combination from the previous cycle is reserved as an elite individual, including: The entire population is divided into several independent islands, and selection, crossover, and mutation operations are performed independently on each island. At the beginning of each control cycle, the target screen parameter combination calculated in the previous cycle is input as an elite individual into the population of the current cycle, and parallel evolutionary calculation of all islands is started.

8. The control system for an efficient and environmentally friendly aggregate screening plant according to claim 1, wherein, The solution process satisfies: when the screening device is a cross screen, the cooperative vibration phase spectrum required to form an interference wave field is a constraint condition, including: For the two exciters of the cross screen, the vibration phase difference satisfies a relationship that is directly proportional to the vibration frequency and the center distance between the two exciters, and inversely proportional to the propagation speed of the vibration wave on the screen surface. During the fitness function evaluation stage of the multi-island genetic algorithm, for individuals that do not satisfy the phase difference constraint, a penalty term is applied to the fitness value, and the value of the penalty term is directly proportional to the absolute difference between the phase difference of the individual solution and the required phase difference.

9. The control system for an efficient and environmentally friendly aggregate screening plant according to claim 1, wherein, When the second derivative of the particle size distribution of the outlet finished product exceeds a preset unstable threshold, the population diversity is enhanced by increasing the mutation probability or inputting random individuals, including: The particle size distribution curve of the outlet finished product is monitored in real time, and the maximum value of the second derivative is calculated. If the maximum value exceeds a preset unstable threshold, the mutation probability is increased from a lower base value to a higher value in the next multi-island genetic algorithm iteration, and a number of completely randomly generated new individuals are forced to be input in each island in each generation to jump out of the local optimal solution.

10. The control system for an efficient and environmentally friendly aggregate screening plant according to any one of claims 1-9, characterized in that, According to the obtained target screen parameter combination, the screen parameters of the screening device are adjusted in real time, including: The target screen parameter combination obtained by optimization is decomposed into specific control instructions, and the target screen parameter combination includes the vibration frequency, amplitude, screen surface inclination angle, and vibration phase difference of the cross screen. PWM signals are output to the frequency converter through the programmable logic controller to adjust the rotational speed of each exciter motor to control the vibration frequency and vibration phase difference. The amplitude is adjusted by adjusting the relative position of the eccentric block. The support angle of the screen box is adjusted by driving the servo motor through the connecting rod mechanism to realize real-time adjustment of the screen surface inclination angle.

Images