Dust fall spraying system for building construction site

Abstract

The invention discloses a building construction site dust fall spraying system, which comprises a distributed three-dimensional perception network, an edge intelligent decision controller and a heterogeneous collaborative execution cluster, and is characterized in that the edge intelligent decision controller comprises a field reconstruction module, a dynamic prediction module and a collaborative optimization module; the heterogeneous cooperative execution cluster comprises a fixed dust suppression curtain wall and a movable dust suppression platform. Through a distributed three-dimensional perception network, real-time and accurate acquisition of multi-dimensional data of a construction environment is realized, and a comprehensive perception capability is provided for the system; the edge intelligent decision controller forms an intelligent brain of the system, can predict the dust diffusion situation in real time based on a physical model, and generates a global optimal collaborative scheduling strategy through a multi-objective optimization algorithm, so that the dust falling operation is converted from passive response to active prediction and accurate planning; therefore, the optimal dust suppression effect is achieved with the minimum resource consumption, dust diffusion is effectively restrained, and meanwhile the automation level and reliability of operation are improved.

Description

Technical Field

[0001] This invention relates to the field of building construction technology, and in particular to a dust suppression spraying system for building construction sites. Background Technology

[0002] At construction sites, to control the large amounts of dust generated by earthwork excavation and material handling, a common dust suppression method is "fixed sprinkler system + mobile mist cannon." Specifically, existing technology typically involves installing a fixed sprinkler system consisting of water supply pipes and nozzles on the perimeter fence of the construction site, supplemented by mobile mist cannons that can be operated manually or with simple remote control. This technical solution aims to cover potential dust-generating areas through pre-set spraying or mobile spraying, replacing the inefficient method of relying entirely on manual watering and meeting basic environmental regulatory requirements.

[0003] However, existing fixed facilities and mobile equipment are isolated from each other, lacking intelligent collaborative decision-making and precise execution capabilities based on the dynamic evolution of global dust, resulting in delayed, inefficient, and resource-wasting dust suppression operations. Fixed sprinkler systems often can only be manually activated or deactivated at set times or in designated areas, unable to adaptively adjust according to real-time dust distribution and wind direction; mobile fog cannons rely on manual observation and operation, resulting in slow response times and a lack of coordination between their operation and that of fixed sprinklers. This prevents precise intervention in the early stages of dust generation and diffusion, often leading to a passive situation where "large-scale spraying only occurs after dust has already spread," making it difficult to effectively curb dust overflow and hindering the efficient use of water and electricity resources. Summary of the Invention

[0004] The purpose of this invention is to provide a dust suppression spraying system for construction sites to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A dust suppression spray system for construction sites includes: The distributed three-dimensional sensing network consists of a dust sensor array deployed at key nodes in the three-dimensional space of the construction site, a miniature meteorological monitoring station deployed within the site, and video monitoring units installed at high points of tower cranes and fences. It is used to collect multi-source heterogeneous data on dust concentration, wind direction and speed, and construction activities in real time. An edge intelligent decision controller, which is communicatively connected to the distributed three-dimensional sensing network, is integrated into an industrial edge computing server at the construction site. The edge intelligent decision controller includes: The field reconstruction module is used to perform spatiotemporal alignment and fusion of the multi-source heterogeneous data, and to construct a continuous dust concentration field and a three-dimensional wind field through a spatial interpolation algorithm. The dynamic prediction module has an embedded numerical model based on the convection-diffusion equation, which is used to predict the dust diffusion trend in the future time domain based on the continuous field data. The collaborative optimization module is used to solve for the optimal collaborative scheduling strategy based on the predicted situation, with multiple objectives including dust suppression effect, resource consumption and operational safety. A heterogeneous collaborative execution cluster, which is communicatively connected to the edge intelligent decision controller, the heterogeneous collaborative execution cluster includes: The fixed dust suppression curtain wall consists of a main water supply pipe laid along the construction boundary, multiple independently controlled regional solenoid valves, and an array of atomizing nozzles on the corresponding branch pipes, used to establish a segmented and controllable basic dust suppression barrier. The mobile dust suppression platform features a chassis capable of moving across complex terrain at construction sites, a high-precision positioning and navigation module, environmental perception and obstacle avoidance sensors, and a directional atomizing cannon mounted via a two-axis servo gimbal.

[0006] Based on the above technical solution, the present invention can be further improved as follows.

[0007] Furthermore, the edge intelligent decision controller is used to analyze the opening and closing commands of the solenoid valves in a specific area, as well as the coordinated control commands for the movement trajectory of the mobile dust suppression platform, the attitude of the atomizing cannon, and its start and stop, driving the two to perform spatiotemporally complementary dust suppression operations.

[0008] Furthermore, the governing equation of the numerical model in the dynamic prediction module, based on the convection-diffusion equation correction, is as follows: in, For dust concentration field, For the three-dimensional wind field, Let be the turbulent diffusion coefficient tensor, and the source term be... The type and location of construction activities identified by the video monitoring unit are calibrated in real time.

[0009] Furthermore, the spatial interpolation algorithm used in the field reconstruction module is Kriging interpolation, which is used to estimate the points... concentration value The estimate is based on the surrounding sensor points. Measured values The linear weighted sum is obtained, and its algorithm formula is as follows: Among them, the weighting coefficient By solving the Kriging equations based on the semi-variogram model, it was determined that the equations satisfy the conditions of unbiasedness and minimization of the estimated variance.

[0010] Furthermore, the collaborative optimization module is used to solve for the optimal collaborative scheduling strategy, and its multi-objective optimization function formula is as follows: in, To predict the concentration field, For safe concentration thresholds, For the total estimated water consumption, The total scheduling time for the mobile platform These are configurable positive weighting coefficients.

[0011] Furthermore, in the fixed dust suppression curtain wall, the spray angle of the atomizing nozzles of the adjacent area branch pipes is set to partially overlap to ensure that the dust suppression barrier remains continuous at the boundary when any adjacent solenoid valves are switched; the area solenoid valve is a normally closed pulse solenoid valve, which receives the pulse width modulation signal sent by the edge intelligent decision controller to control the opening degree and duration.

[0012] Furthermore, the environmental perception and obstacle avoidance sensors of the mobile dust suppression platform include a forward-facing lidar and a stereo vision camera; the high-precision positioning and navigation module integrates GNSS, inertial measurement unit and wheel odometer data; the two-axis servo gimbal is used to continuously rotate in the horizontal plane and pitch and yaw in the vertical plane, and its motion is set by the target pitch angle and yaw angle in the cooperative control command.

[0013] Furthermore, the edge intelligent decision controller also includes a digital twin verification and simulation platform. Before the control commands from the collaborative optimization module are sent to the physical execution cluster, the digital twin verification and simulation platform performs advanced simulation by loading a virtual scene containing a 3D model of the construction site, an equipment model, and collaborative control commands. This is used to detect collisions in the movement path and interference risks in the work area, and the verification results are fed back to optimize the command sequence.

[0014] Furthermore, the edge intelligent decision controller also includes an online self-learning module, which collects historical prediction data, scheduling instructions, and subsequent actual dust concentration data; calculates prediction error and strategy effectiveness indicators; and dynamically adjusts key parameters in the dynamic prediction model based on the prediction error using a parameter estimation algorithm. Based on the aforementioned strategy performance index, the weight coefficients in the collaborative optimization module are dynamically adjusted using an optimization algorithm. .

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention achieves real-time and accurate acquisition of multi-dimensional data on the construction environment through a distributed three-dimensional perception network composed of a dust sensor array, a weather station, and a video monitoring unit, providing the system with comprehensive perception capabilities and overcoming the limitations of existing technologies that rely on manual observation or isolated point data. Based on this, the edge intelligent decision controller and its embedded field reconstruction, dynamic prediction, and collaborative optimization modules constitute the system's "intelligent brain," capable of predicting dust diffusion trends in real time based on physical models and generating globally optimal collaborative scheduling strategies through multi-objective optimization algorithms, thereby transforming dust suppression operations from passive response to proactive prediction and precise planning. Simultaneously, the heterogeneous collaborative execution cluster composed of zoned controllable fixed dust suppression curtain walls and autonomously navigating mobile dust suppression platforms, along with its architecture tightly connected to the intelligent decision-maker, enables precise execution of the global optimization strategy. This achieves seamless spatiotemporal complementarity and collaboration between the on-demand, segmented, precise opening and closing of fixed barriers and the autonomous approach and suppression by mobile platforms, completely changing the isolated operation mode between equipment. Thus, it achieves optimal dust suppression with minimal resource consumption, effectively curbing dust diffusion while improving the level of automation and reliability of operations. Attached Figure Description

[0016] Figure 1 This is a schematic block diagram of the hardware structure of the dust suppression spraying system for construction sites according to the present invention; Figure 2 This is a schematic diagram of the edge intelligent decision controller functional module and decision process of the dust suppression spray system at the construction site of the present invention; Figure 3 This is a flowchart illustrating the learning and optimization process of the online self-learning module of the dust suppression spraying system at the construction site of this invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings.

[0018] This invention provides a dust suppression spraying system for construction sites, whose hardware and software architecture work together to achieve intelligent dust suppression. For example... Figure 1 As shown, a dust suppression spraying system for construction sites consists of three main parts.

[0019] Distributed three-dimensional sensing network: This distributed three-dimensional sensing network uses dust sensors (such as PM2.5 / PM10 sensors based on laser scattering principle) arranged in a three-dimensional spatial grid within the construction site, especially adding monitoring points near conventional ground monitoring points, material storage yards, main roads, and construction layers on the exterior facades of high-rise buildings, to capture the vertical distribution of dust; it also deploys miniature meteorological monitoring stations in the center of the site and upwind to measure wind speed, wind direction, temperature, and humidity in real time; and it installs high-definition PTZ cameras and bullet cameras under the tower crane cab and at the highest point of the perimeter of the construction site to form a comprehensive video surveillance coverage without blind spots, used to automatically identify dust activities such as earthwork excavation, vehicle movement, and building material loading and unloading.

[0020] Edge Intelligent Decision Controller: The edge intelligent decision controller uses an industrial-grade edge computing server deployed in the power distribution room or command center at the construction site as its carrier. All sensing data is aggregated to the edge intelligent decision controller via industrial Ethernet or 5G CPE. The edge intelligent decision controller runs decision-making software that integrates a field reconstruction module, a dynamic prediction module, and a collaborative optimization module. The field reconstruction module is used to perform spatiotemporal alignment and fusion of the multi-source heterogeneous data, and construct a continuous dust concentration field and a three-dimensional wind field through spatial interpolation algorithms. The dynamic prediction module has an embedded numerical model based on the convection-diffusion equation correction, which is used to predict the dust diffusion trend in the future time domain based on the continuous field data. The collaborative optimization module is used to solve for the optimal collaborative scheduling strategy based on the predicted trend, with dust suppression effect, resource consumption, and operational safety as multiple objectives.

[0021] Heterogeneous Collaborative Execution Cluster: The heterogeneous collaborative execution cluster communicates with the edge intelligent decision controller via a wireless local area network. The heterogeneous collaborative execution cluster includes a fixed dust suppression curtain wall and a mobile dust suppression platform. The main water supply pipe of the fixed dust suppression curtain wall is laid along the entire construction enclosure. A control zone is set every 20-30m. Each zone is controlled by a normally closed pulse solenoid valve. Fan-shaped atomizing nozzles with an atomization angle of 90-120° are installed on its branch pipes. The spraying range of the nozzles in adjacent zones is designed to overlap by about 15%-20%. The mobile dust suppression platform includes a tracked chassis. The tracked chassis is equipped with an integrated GNSS / INS navigation system, a lidar, and a binocular stereo vision camera for environmental perception and obstacle avoidance sensors. The tracked chassis is also equipped with a two-axis servo gimbal (yaw axis ±180°, pitch axis -10° to +45°) to support the atomizing cannon, which is used to adjust the attitude of the atomizing cannon.

[0022] In a preferred embodiment, the present invention may be further configured as follows: Figure 2 As shown in the figure, this embodiment specifically illustrates the core algorithm model of the dynamic prediction module, which is used to convert the continuous three-dimensional wind field output by the field reconstruction module into a dynamic prediction module. Using the dust concentration field as input, its core is a simplified unsteady convection-diffusion equation numerical model, and the governing equation is: in, Represents spatial location and time Dust concentration (unit: μg / m³); It is the wind speed vector field (m / s); It is the turbulent diffusion coefficient tensor. Under the simplified assumption of horizontal isotropy, the diffusion coefficients in the horizontal and vertical directions can be mainly considered. and ; source item As a key innovation, it dynamically calibrates and maps the type (e.g., excavator, dump truck), location, and operational intensity (e.g., estimated based on visual characteristics of dust) of construction machinery identified in real time by the video monitoring unit using image recognition algorithms (such as YOLO). For example, it identifies a construction machine located in... If an excavator is performing earthmoving operations, then a source strength source, lasting for several minutes and with an intensity related to the operation status, is applied near that location. The model uses the finite volume method for discretization and numerical solution. Taking the current moment as the initial condition, it predicts the spatial distribution evolution of dust concentration in the next 5 to 15 minutes and generates a dynamic situation map.

[0023] In a preferred embodiment, the present invention may be further configured as follows: Figure 2 As shown in this embodiment, the spatial interpolation algorithm used in the field reconstruction module employs Kriging interpolation. Since the sensors are discretely deployed, to obtain a continuous spatial distribution field across the entire construction area, this system uses Kriging interpolation, a widely used optimal unbiased estimation method in geostatistics. For any point to be estimated within the site... (coordinates are) The estimated dust concentration From the surrounding Known sensor measurements The linear weighted sum is obtained, and its algorithm formula is as follows: Weighting coefficient It is not simply determined by distance, but rather obtained by solving a system of Kriging equations. This system is based on a semi-variogram model, which describes how the variance of measurements between two points in space varies with distance (such as the spherical model or the exponential model). The solution process requires two conditions: first, the unbiasedness condition, meaning the sum of all weights is 1. Secondly, the variance minimization condition is estimated. This is achieved by solving the system of equations. Enables the estimated value It is statistically optimal and can reflect the spatial structure characteristics of the data. For wind fields, the wind speed and wind direction components can be interpolated separately.

[0024] In a preferred embodiment, the present invention may be further configured as follows: Figure 2 As shown, in this embodiment, the decision-making mechanism of the collaborative optimization module receives the concentration field of future time periods output by the dynamic prediction module. Its decision-making objective is not singular, but needs to strike a balance between dust suppression effect, water consumption, and equipment scheduling efficiency; therefore, the following multi-objective optimization function is constructed. The solution is given by the following function formula: in, This is a dust suppression effect item, and its physical meaning is to calculate and predict the concentration field exceeding the safe threshold. (For example, the volume of a region with a PM10 concentration of 75 μg / m³) (integral with respect to the concentration) that the system needs to minimize is this "excess volume"; It is a resource consumption item, which is proportional to the estimated total water consumption of all fixed and mobile units during the strategy execution cycle. It is an efficiency penalty item, mainly related to the total scheduling distance or time of the mobile platform, used to encourage efficient task allocation. This is a positive weight coefficient that the administrator can dynamically adjust based on site environmental protection requirements, water resource costs, and operational urgency. The optimization problem is solved once per decision cycle (e.g., 1 minute). The decision variables are the set of working states of all execution units (the opening and closing states of each solenoid valve, the target pose of the moving platform, and operational parameters). Heuristic algorithms such as genetic algorithms and particle swarm optimization can be used to solve the problem, ultimately outputting a value that satisfies the objective function. The smallest "optimal cooperative scheduling strategy".

[0025] In a preferred embodiment, the present invention may be further configured as follows: Figure 1As shown, the main water supply pipe for the fixed dust suppression curtain wall typically uses DN50 or DN80 PE pipes, laid along the inner side of the construction site enclosure or slightly above the ground. The branch pipes for each zone use DN25 PE pipes, extending vertically upwards from the main pipe to the top of the enclosure, and then horizontally extending along the top of the enclosure for a short distance. Each zone is independently controlled by a normally closed pulse solenoid valve (e.g., DC24V, DN25). In this embodiment, the edge intelligent decision controller sends a pulse width modulation (PWM) signal. By adjusting the duty cycle of the PWM signal, the average opening degree of the solenoid valve within one cycle can be precisely controlled, thereby achieving stepless or step-wise adjustment of the water volume. For example, a small water volume spray can be used when the dust concentration is low, and the valve can be fully opened when the concentration is high. Each zone branch pipe is equipped with 6 to 8 fan-shaped atomizing nozzles. The installation angle is precisely calculated to ensure that the coverage area of ​​the nozzles in this zone overlaps with that of the adjacent zone by about 15 to 20%. This design ensures that even if only one zone is activated, an effective local barrier can be formed; and when adjacent zones need to take over, the barrier is continuous and without gaps at the boundary.

[0026] In a preferred embodiment, the present invention may be further configured as follows: Figure 1 As shown, the mobile dust suppression platform is based on an off-road capable electric or diesel-powered tracked chassis. Its core navigation and positioning module employs multi-sensor fusion technology, combining GNSS, inertial measurement unit (IMU), and wheeled odometer. GNSS provides absolute position (accuracy decreases in signal obstruction areas), the IMU provides continuous attitude and acceleration information, and the wheeled odometer provides relative displacement. By fusing these three types of data through a Kalman filter, centimeter- to decimeter-level positioning accuracy can be maintained even under brief building obstructions to the GNSS signal. Environmental perception and obstacle avoidance sensors are used to ensure mobile... The dust suppression platform moves safely and autonomously. The lidar is responsible for constructing a 3D point cloud map within a 50m range in front of the mobile dust suppression platform to detect the outlines and distances of static and dynamic obstacles. The binocular stereo vision camera assists in the identification of obstacle types (such as distinguishing between workers, piles of building materials, and pits). The two-axis servo gimbal carries the atomizing cannon, whose flow rate is adjustable from 50 to 200 L / min. The two-axis servo gimbal receives target angle commands (yaw angle, pitch angle) from the edge intelligent decision controller and quickly and accurately drives the cannon to aim at the target dust source area through closed-loop servo control.

[0027] In a preferred embodiment, the present invention may be further configured as follows: Figure 2As shown, the digital twin verification and simulation platform, as an independent module of the decision-making software in the edge intelligent decision controller, maintains a virtual 3D scene in memory that is synchronized 1:1 with the physical construction site. This scene includes not only static terrain, building, and fence models, but also dynamically updates the real-time position and speed of each mobile platform and the on / off status of fixed equipment. After the collaborative optimization module generates a preliminary scheduling strategy and compiles it into a low-level control instruction sequence, these low-level control instructions are not directly issued. Instead, they are first input into the digital twin verification and simulation platform for "advanced simulation." The platform runs the simulation at ultra-real-time speed, simulating the movement of mobile platforms according to instructions, the rotation of the two-axis servo gimbal, the activation of fog cannons, and the operation of fixed sprinkler valves. The entire opening and closing process involves the system automatically detecting: a) whether the planned path of the mobile platform collides with static obstacles (such as buildings or pits) or the predicted paths of other mobile platforms or construction machinery; b) whether the atomized jet range of the mobile platform will spatially interfere with the high-altitude hoisting area or high-voltage power lines currently in operation; and c) whether the operating ranges of different mobile platforms overlap, leading to resource conflicts. Once any conflict is detected, the digital twin verification and simulation platform immediately feeds back the conflict type and location information to the collaborative optimization module. Based on the feedback, the collaborative optimization module replans and generates a new instruction sequence within seconds, and performs simulation verification again until it passes. This "simulation-optimization" iterative cycle greatly improves the safety and reliability of the entire system operation.

[0028] In a preferred embodiment, the present invention may be further configured as follows: Figure 3 As shown, the online self-learning module serves as the core of the system's continuous optimization. This module continuously collects and stores a closed-loop data chain. Its learning process unfolds in parallel at two levels. At the model calibration level, the online self-learning module periodically (e.g., every 24 hours) compares the differences between historical predicted concentration fields and actual concentration fields, calculates the regional average error, and automatically adjusts key parameters in the dynamic prediction model using methods such as gradient descent or Bayesian inference. The (turbulent diffusion coefficient) parameter allows the model's predicted output to continuously approximate reality, thus adapting to changes in diffusion characteristics across different seasons and construction site layouts. At the strategy optimization level, the module evaluates the "cost-effectiveness" of historical scheduling strategies, comprehensively considering their dust suppression effect (actual reduction in excessive volume) and the associated water and electricity costs. Based on this evaluation data, the online self-learning module can utilize reinforcement learning (such as Q-learning) or evolutionary algorithms to dynamically fine-tune the multi-objective function in the collaborative optimization module. Weighting coefficients For example, automatically increasing [efforts] during the dry season. (Water conservation weight), or automatically increased during environmental inspections. (Dust suppression effect weight) enables the system's decision preferences to adapt to changes in external constraints.

[0029] The control method of the dust suppression spray system at a construction site according to the present invention specifically includes the following steps: S1. During operation, the various sensors and cameras in the distributed three-dimensional perception network work continuously to transmit the collected dust concentration, meteorological data and video stream to the edge intelligent decision controller in real time through wired or wireless networks. S2. Next, the field reconstruction module first performs timestamp alignment and formatting on all collected data, and then calls the Kriging interpolation algorithm to reconstruct the discrete sensor point data into three-dimensional continuous concentration field and wind field grid data covering the entire construction area. S3. The dynamic prediction module loads the latest continuous field data and the dynamic dust source term generated by the video analysis results, substitutes it into its built-in modified convection-diffusion equation model for numerical solution, and outputs a prediction map of the spatial evolution of dust concentration in the future (e.g., 10 min). S4. The collaborative optimization module reads the prediction map, combines it with the current execution unit status, constructs and solves a multi-objective optimization problem with dust suppression, water conservation and high efficiency as the core, and generates an optimal collaborative scheduling strategy that specifies in detail "which fixed partition is switched on and off when, and where the mobile platform goes to work". This high-level scheduling strategy is parsed by the resource scheduler into a set of specific, time-ordered preliminary control instructions. S5. By sending the initial control commands to the digital twin verification platform for advanced simulation, only the command sequence that passes all collision and conflict detection can proceed to the next step. The verified command sequence is sent to the solenoid valve controllers of the fixed dust suppression curtain wall and the on-board computer of the mobile dust suppression platform through a reliable communication link, driving the fixed dust suppression curtain wall and the mobile dust suppression platform to strictly coordinate their actions according to the commands and complete the precise dust suppression operation. S6. After completing the execution of the operation, the sprinkler system immediately enters the learning and optimization phase. At this time, the distributed three-dimensional sensing network continues to work, monitoring the actual dust concentration field changes after the dust suppression operation (e.g., 5-10 minutes) and transmitting this as "real effect" data. The online self-learning module then starts, comparing the actual concentration field at this moment with the predicted concentration field generated in the dynamic prediction module at the same time point by point, and calculating the global average absolute error or root mean square error as the model prediction error. At the same time, combined with the recorded resource consumption data, the actual performance indicators of the executed scheduling strategy (such as the dust suppression efficiency per unit of water consumption) are evaluated. Then, these errors and performance data are used to drive two optimization processes in parallel: First, the model prediction error is used as a feedback signal to automatically adjust the key parameters affecting the accuracy of the dynamic prediction model, such as the turbulent diffusion coefficient, through parameter estimation algorithms (such as recursive least squares method). First, to make the model's next prediction more accurate; second, to determine the weight coefficients in the current multi-objective optimization based on the actual performance indicators of the strategy. If the expected balance is not achieved, the system is fine-tuned using an optimization algorithm. After these adjustments, the updated model parameters and decision weights are applied to the next control cycle, enabling the entire system to continuously learn from historical experience and improve itself. This achieves true closed-loop adaptive optimization and ensures the long-term effectiveness and economy of the system.

[0030] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but all such changes and modifications fall within the scope of protection of the present invention.

Claims

1. A dust suppression spray system for a construction site, characterized in that, include: The distributed three-dimensional sensing network includes a dust sensor array deployed at the construction site, a miniature meteorological monitoring station deployed within the site, and video monitoring units installed at high points on tower cranes and fences, used to collect multi-source heterogeneous data on dust concentration, wind direction and speed, and construction activities in real time. An edge intelligent decision controller, which is communicatively connected to the distributed three-dimensional sensing network, is integrated into an industrial edge computing server at the construction site. The edge intelligent decision controller includes: The field reconstruction module is used to perform spatiotemporal alignment and fusion of the multi-source heterogeneous data, and to construct a continuous dust concentration field and a three-dimensional wind field through a spatial interpolation algorithm. The dynamic prediction module has an embedded numerical model based on the convection-diffusion equation, which is used to predict the dust diffusion trend in the future time domain based on the continuous field data. The collaborative optimization module is used to solve for the optimal collaborative scheduling strategy based on the predicted situation, with multiple objectives including dust suppression effect, resource consumption and operational safety. A heterogeneous collaborative execution cluster, which is communicatively connected to the edge intelligent decision controller, the heterogeneous collaborative execution cluster includes: A fixed dust suppression curtain wall includes a main pipe, several branch pipes, multiple independently controlled regional solenoid valves, and atomizing nozzle arrays on the corresponding branch pipes, used to establish a segmented and controllable basic dust suppression barrier. The mobile dust suppression platform includes a mobile chassis and a high-precision positioning and navigation module, an environmental perception and obstacle avoidance sensor, a two-axis servo gimbal, and an atomizing cannon mounted on the mobile chassis. The atomizing cannon is mounted on the two-axis servo gimbal and is used to drive the attitude adjustment of the atomizing cannon.

2. A dust suppression spray system for a construction site according to claim 1, wherein, The edge intelligent decision controller is connected to the heterogeneous collaborative execution cluster for parsing the opening and closing commands of the solenoid valves in a specific area, as well as the collaborative control commands for the movement trajectory, atomizing cannon posture and start / stop of the mobile dust suppression platform, driving the mobile dust suppression platform and the fixed dust suppression curtain wall to perform spatiotemporally complementary dust suppression operations.

3. A dust suppression spray system for a construction site according to claim 1, wherein, The governing equations of the numerical model based on the convection-diffusion equation correction in the dynamic prediction module are as follows: in, For dust concentration field, For the three-dimensional wind field, Let be the turbulent diffusion coefficient tensor, and the source term be... The type and location of construction activities identified by the video monitoring unit are calibrated in real time.

4. A dust suppression spraying system for construction sites according to claim 1, characterized in that, The spatial interpolation algorithm used in the field reconstruction module is Kriging interpolation, which is used to estimate the points... concentration value The estimate is based on the surrounding sensor points. Measured values The linear weighted sum is obtained, and its algorithm formula is as follows: where the weight coefficient The Kriging equation set based on the semi-variogram model is determined by solving, which satisfies the unbiasedness and the minimum estimation variance conditions.

5. A dust suppression spray system for a construction site according to claim 1, wherein, The collaborative optimization module is used to solve for the optimal collaborative scheduling strategy, and its multi-objective optimization function formula is as follows: wherein, is a predicted concentration field, is a safety concentration threshold, is a total estimated water consumption, is a total scheduled time of movement platform, is a configurable positive weight coefficient.

6. A dust suppression spray system for a construction site according to claim 1, wherein, In the fixed dust suppression curtain wall, the spray angle of the atomizing nozzles of the adjacent area branch pipes is set to partially overlap to ensure that the dust suppression barrier remains continuous at the boundary when any adjacent solenoid valves are switched; the area solenoid valve is a normally closed pulse solenoid valve, which receives the pulse width modulation signal sent by the edge intelligent decision controller to control the opening degree and duration.

7. A dust suppression spray system for a construction site according to claim 1, wherein, The environmental perception and obstacle avoidance sensors of the mobile dust suppression platform include a forward-facing lidar and a stereo vision camera; the high-precision positioning and navigation module integrates GNSS, inertial measurement unit and wheel odometer data; the two-axis servo gimbal is used to rotate continuously in the horizontal plane and pitch and yaw in the vertical plane, and its motion is set by the target pitch angle and yaw angle in the cooperative control command.

8. A dust suppression spray system for a construction site according to claim 1, wherein, The edge intelligent decision controller also includes a digital twin verification and simulation platform. Before the control commands from the collaborative optimization module are sent to the physical execution cluster, the digital twin verification and simulation platform performs advanced simulation by loading a virtual scene containing a 3D model of the construction site, an equipment model, and collaborative control commands. This is used to detect collisions in the movement path and interference risks in the work area, and the verification results are fed back to optimize the command sequence.

9. A dust suppression spraying system for construction sites according to claim 1, characterized in that, The edge intelligent decision controller further comprises an online self-learning module, which is used for collecting prediction data, scheduling instructions and subsequent actual dust concentration data of a historical period; calculating a prediction error and a strategy performance index; and dynamically adjusting key parameters in the dynamic prediction model based on the prediction error through a parameter estimation algorithm ​ Based on the aforementioned strategy effectiveness index, the weight coefficients in the collaborative optimization module are dynamically adjusted using an optimization algorithm. .

Images