Construction site flying dust noise monitoring system

Abstract

The invention discloses a construction site dust and noise monitoring system, which relates to the technical field of environment monitoring, and comprises a multi-mode sensing fusion unit, a dust monitoring unit, a noise monitoring unit, a laser scattering sensor, a beta-ray sensor, a microphone array sound source orientation and vibration sensor spectrum analysis unit and a noise detection unit, by combining environmental parameters such as temperature and humidity, air pressure and raindrops, a multi-dimensional sensing network is constructed, and the pain point of insufficient reliability of a single data source of traditional equipment is solved; dynamic environment adaptability is achieved, the humidity compensation unit corrects errors of the laser sensor in a segmented mode according to real-time temperature and humidity, the noise separation unit filters out non-construction interference through voiceprint-vibration spectrum matching, the drift correction unit automatically calibrates the laser sensor based on a beta-ray absolute value, and the data credibility of a high-humidity and multi-interference scene is remarkably improved; targeted intelligent treatment is conducted, when PM10 continuously exceeds the standard and noise exceeds the limit, the system triggers positioning spraying and directional sound wave suppression, and meter-scale precision pollution control is achieved through sound-light alarm and work order pushing.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, specifically to a construction site dust and noise monitoring system. Background Technology

[0002] With the promulgation of documents such as the Air Pollution Prevention and Control Law and the Technical Standard for Online Monitoring of Dust from Construction Projects (T / CECS1019-2022), many regions have mandated the installation of dust and noise monitoring equipment at construction sites. For example, Nanjing stipulates that projects with a construction cost exceeding 100 million yuan must be equipped with data visualization monitoring equipment, while Shenzhen has achieved full-element supervision through the "Remote Stop 3.0" system.

[0003] The limitations of traditional monitoring methods include low efficiency of manual inspections, reliance on regular manual testing, and inability to respond to pollution incidents in real time; data silos, with scattered monitoring equipment lacking a unified platform for integration, leading to fragmented supervision; and outdated treatment methods, with traditional sprinkler systems relying on manual start-up and shutdown, unable to be dynamically adjusted based on real-time data.

[0004] The maturity of technologies such as 5G, edge computing, and AI has made it possible to build an integrated "air-space-ground" monitoring network. For example, Bonn Instruments' system uses video overlay technology to achieve three-dimensional monitoring of "data + images," reducing the duration of dust exceeding standards by 82%.

[0005] Therefore, there is an urgent need for a construction site dust and noise monitoring system to solve the problems of insufficient sensor accuracy and environmental adaptability, including the many interference factors, the susceptibility of laser sensors to water mist in high humidity environments, the potential interference of noise sensors with background noise (such as traffic and machinery noise) leading to data distortion; and the complexity of calibration, as traditional sensors require regular manual calibration, while the harsh environment of construction sites (such as high temperature and high dust) accelerates equipment aging and increases maintenance costs. Summary of the Invention

[0006] To address the aforementioned technical problems, a construction site dust and noise monitoring system is provided. This technical solution solves the problems of insufficient sensor accuracy and environmental adaptability.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A construction site dust and noise monitoring system includes:

[0009] Multimodal sensing module: Dust monitoring unit, integrating a laser scattering sensor and a beta-ray sensor, wherein the laser scattering sensor collects PM2.5 / PM10 concentration in real time, and the beta-ray sensor is calibrated according to a preset period; Noise monitoring unit, deploying a microphone array and a vibration sensor, wherein the microphone array uses acoustic beamforming technology to directionally capture sound sources in the construction area, and the vibration sensor simultaneously collects the mechanical vibration spectrum to separate background noise; Environmental parameter unit, configured with temperature and humidity sensors, barometer and raindrop sensor;

[0010] Environmental adaptive calibration module: Humidity compensation unit, which dynamically corrects the water mist interference error of the laser scattering sensor based on temperature and humidity data. When the humidity is greater than or equal to a predetermined ratio, it automatically switches to the output dominated by the β-ray sensor; Noise separation unit, which filters out non-construction-related traffic noise and mechanical resonance interference by comparing the spectrum of the vibration sensor with the acoustic characteristics of the microphone array; Drift correction unit, which performs automatic zero-point calibration of the laser scattering sensor at regular intervals using the absolute measurement value of the β-ray sensor.

[0011] Intelligent response linkage module: connects the spray controller and the sound and light alarm. When the PM10 concentration exceeds the threshold for a set number of minutes or the noise exceeds the standard, it triggers the positioning spray and directional sound wave suppression.

[0012] Preferably, the multimodal sensing module specifically includes:

[0013] Dust monitoring unit:

[0014] A laser scattering sensor irradiates suspended particulate matter in the air with a laser beam of a specific wavelength and calculates the concentration values ​​of PM2.5 and PM10 in real time based on the intensity distribution of the scattered light from the particles.

[0015] A beta-ray sensor uses an internal radioactive source to release beta rays that penetrate a sampling filter membrane, and the mass concentration of particulate matter is accurately determined by the amount of ray attenuation.

[0016] The laser scattering sensor continuously outputs monitoring data, while the beta ray sensor automatically initiates the calibration process at regular intervals, feeding back the absolute measurement value to the processing core to dynamically correct the zero-point drift and range error of the laser sensor.

[0017] Preferably, the multimodal sensing module specifically includes:

[0018] Noise monitoring unit:

[0019] The microphone array consists of high-sensitivity digital microphones arranged in a diamond topology. It uses a beamforming algorithm to spatially filter the sound field in the construction area and generate a directional pickup beam with an adjustable angle range.

[0020] The vibration sensor uses a MEMS triaxial accelerometer to collect the vibration spectrum characteristics within a predetermined wide frequency range;

[0021] The noise separation engine performs time-frequency domain matching between the sound signal and the vibration signal. When the microphone array captures a sound pressure event, it searches the synchronous vibration spectrum database. If a mechanical equipment characteristic spectrum that is time-aligned with the sound event is detected, it is determined to be valid construction noise. If the sound event has no matching vibration characteristics, it is regarded as background interference and is eliminated.

[0022] Preferably, the multimodal sensing module specifically includes:

[0023] Environmental parameter unit:

[0024] The temperature and humidity sensor uses a capacitive polymer thin film probe to monitor the ambient temperature and humidity within a predetermined temperature range, providing a humidity correction coefficient for dust monitoring.

[0025] The barometer is based on a piezoresistive MEMS chip and combines temperature data to calculate air density, correcting the gas absorption effect error of the β-ray sensor.

[0026] The rain sensor detects precipitation intensity through an infrared optical grid. When a raindrop hits the monitoring window, it triggers a rainproof mode, closes the air intake valve of the laser scattering sensor to reduce the risk of wet pollution, and activates a sedimentation compensation model for dust caused by rainfall.

[0027] Preferably, the environment adaptive calibration module specifically includes:

[0028] Humidity compensation unit:

[0029] The real-time correction mechanism involves the temperature and humidity sensor collecting environmental data every second. When the relative humidity is less than or equal to a predetermined ratio, the humidity compensation algorithm is activated, and the correction coefficient is calculated in segments based on the humidity value. When the humidity value is greater than the predetermined ratio, the output is automatically switched to the β-ray sensor, and the laser scattering data is used as an auxiliary reference.

[0030] Based on the status of the raindrop sensor, if continuous precipitation is detected, the β-ray dominant mode will be retained even if the humidity value is greater than a predetermined ratio; if there is only high humidity but no precipitation, the mode will be switched back to laser dominant mode after the humidity drops back to the predetermined ratio.

[0031] Preferably, the environment adaptive calibration module specifically includes:

[0032] Noise separation unit:

[0033] Synchronous signal acquisition: a microphone array captures sound pressure signals at a predetermined sampling rate, and a vibration sensor synchronously acquires triaxial acceleration data at the contact points of the construction machinery.

[0034] The interference filtering process includes feature extraction, which includes acoustic signals to extract typical acoustic signature features of construction equipment and vibration signals to identify the fundamental frequency harmonics of the equipment.

[0035] Correlation determination involves calculating the cross-correlation coefficient between the acoustic signal and the vibration signal within a time window. If the cross-correlation coefficient is greater than or equal to a predetermined threshold and the frequency characteristics match, it is determined to be valid construction noise.

[0036] Dynamic threshold adjustment adapts to background noise intensity at different times, based on the ambient noise background value.

[0037] Preferably, the environment adaptive calibration module specifically includes:

[0038] Drift correction unit:

[0039] The timed calibration is triggered, and the calibration sequence is started at a set time. The β-ray sensor automatically collects particulate matter mass concentration data and records the laser scattering sensor readings at the same time.

[0040] Drift compensation mechanism, zero-point calibration: when PM10 is detected to be continuously lower than the predetermined threshold during the construction site's quiet period, the baseline value of the laser scattering sensor is forced to zero; range calibration: compare the absolute measurement value of β-rays with the laser scattering reading, generate a proportional coefficient, divide the proportional coefficient range, and trigger calibration strategies in different ranges.

[0041] Self-diagnostic fault tolerance: When the load on the β-ray sensor membrane exceeds a predetermined ratio, calibration is automatically paused, triggering the solar self-cleaning module to remove accumulated dust and ensure the reliability of the calibration reference.

[0042] Preferably, the intelligent response linkage module specifically includes:

[0043] Multi-parameter fusion determination mechanism:

[0044] Dynamic identification of pollution events; triggering conditions for dust exceeding standards: when the PM10 concentration exceeds the preset threshold for a continuous period of time, and the environmental parameter unit excludes precipitation interference, it is determined to be a valid dust event; triggering conditions for noise exceeding standards: when construction noise, after being filtered by the noise separation unit, exceeds the legal limit for a continuous period of time, and the vibration sensor simultaneously verifies the equipment operating status; priority of compound events: if dust and noise exceed standards at the same time, dust suppression measures are initiated first; for single events, a targeted response is triggered.

[0045] Pollution heat maps are generated based on data from multimodal sensing units, including dust source location, combining wind speed and direction sensor data and PM concentration gradient changes to invert the pollution diffusion path; noise source location, calculating the time difference of arrival of sound through microphone array beamforming to pinpoint the coordinates of equipment exceeding the standard.

[0046] Preferably, the intelligent response linkage module specifically includes:

[0047] Precise dust suppression implementation:

[0048] Positioning-based sprinkler control, intelligent sprinkler networking, deployment of IoT smart sprinklers, and automatic adjustment of spray angle after receiving the coordinates of the pollution source;

[0049] Dynamic water volume optimization: for light pollution, intermittent spraying from a single nozzle is activated; for heavy pollution, all nozzles within a predetermined radius are activated to work continuously, and a fog cannon truck is linked to enhance coverage; water-saving strategy: based on feedback from the raindrop sensor, if the rainfall intensity exceeds a predetermined threshold, the spraying is automatically paused to utilize natural rainfall for dust suppression.

[0050] Directional sound wave suppression and sound wave interference noise reduction: For fixed noise sources, reverse sound waves are generated through sound field modeling to form a silencing zone in the target area; dynamic tracking of moving sound sources, combined with UWB positioning module to update the phase of the reverse sound waves in real time; sound and light alarm coordination: flashing LED warnings are triggered around the equipment exceeding the standard, and directional voice prompts are played simultaneously; after a predetermined period of continuous exceeding the standard, an alarm work order is automatically pushed to the mobile phone of the responsible personnel.

[0051] Preferably, the intelligent response linkage module specifically includes:

[0052] Predictive linkage control:

[0053] AI-driven proactive intervention, a dust-noise correlation model, and real-time analysis of construction plan data. Input parameters include equipment start-up and shutdown plans, earthwork operation area, and short-term weather forecasts. Output actions include pre-starting the sprinkler system to wet the work surface before the predicted exceedance.

[0054] Closed-loop performance evaluation involves continuous monitoring of pollution parameter change rates after response. If the PM10 concentration does not decrease as expected within a predetermined timeframe, the spray pressure is automatically increased and additional backup spray nozzles are activated. A response performance report is generated to optimize control parameters for the next operation.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0056] This invention overcomes the limitations of traditional single-sensor equipment by integrating high-frequency laser scattering monitoring and β-ray reference calibration through multimodal sensing; and achieves accurate separation of construction noise through acoustic-vibration coupling technology, solving the problem of false alarms in broadband noise.

[0057] Environmental adaptive evolution capability, intelligent humidity compensation, and a segmented compensation model compress the monitoring error in high humidity environments; closed-loop drift correction, based on an automatic calibration system using the absolute value of β rays, reduces the zero-point drift of the laser sensor within six months; dynamic noise filtering, and spatiotemporal matching technology of acoustic signature and vibration spectrum, improve the success rate of filtering out interference sources such as traffic noise.

[0058] Precise targeted control, combined with wind speed and direction inversion to identify pollution sources, enables water-saving spraying and improves dust suppression efficiency; directional sound wave noise reduction creates a 5-meter silencing zone around equipment exceeding standards, reducing noise by 12dB in specific frequency bands, avoiding the waste of resources in "complete silence".

[0059] The predictive intervention mechanism, which integrates construction plans and meteorological data using an AI model, pre-activates dust suppression 30 minutes in advance, reducing the incidence of dust exceeding standards. Attached Figure Description

[0060] Figure 1 This is an internal framework diagram of a construction site dust and noise monitoring system. Detailed Implementation

[0061] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0062] Reference Figure 1 As shown, a construction site dust and noise monitoring system includes:

[0063] Multimodal sensing module: Dust monitoring unit, integrating a laser scattering sensor and a beta-ray sensor, wherein the laser scattering sensor collects PM2.5 / PM10 concentration in real time, and the beta-ray sensor is calibrated according to a preset period; Noise monitoring unit, deploying a microphone array and a vibration sensor, wherein the microphone array uses acoustic beamforming technology to directionally capture sound sources in the construction area, and the vibration sensor simultaneously collects the mechanical vibration spectrum to separate background noise; Environmental parameter unit, configured with temperature and humidity sensors, barometer and raindrop sensor;

[0064] Environmental adaptive calibration module: Humidity compensation unit, which dynamically corrects the water mist interference error of the laser scattering sensor based on temperature and humidity data. When the humidity is greater than or equal to a predetermined ratio, it automatically switches to the output dominated by the β-ray sensor; Noise separation unit, which filters out non-construction-related traffic noise and mechanical resonance interference by comparing the spectrum of the vibration sensor with the acoustic characteristics of the microphone array; Drift correction unit, which performs automatic zero-point calibration of the laser scattering sensor at regular intervals using the absolute measurement value of the β-ray sensor.

[0065] Intelligent response linkage module: connects the spray controller and the sound and light alarm. When the PM10 concentration exceeds the threshold for a set number of minutes or the noise exceeds the standard, it triggers the positioning spray and directional sound wave suppression.

[0066] It should be noted that the environmental parameter unit generates temperature, humidity, and air pressure matrices in real time, which drives the dust monitoring unit to switch the main sensor (laser / β-ray) and correct the noise separation algorithm to ensure the physical authenticity of the input data.

[0067] When dust and noise events occur simultaneously, the system prioritizes the suppression of the dominant pollution source based on the pollution diffusion model (e.g., priority spraying for PM10 exceeding the standard, priority sound wave suppression for high-frequency noise).

[0068] After the spray system is started, the dust suppression efficiency is assessed by the rate of PM10 concentration decrease, and the spray intensity is dynamically adjusted to form a closed loop of "monitoring → intervention → verification".

[0069] Cross-module redundancy and fault-tolerant design:

[0070] In the event of a laser scattering sensor malfunction, the beta-ray sampling frequency is automatically increased to once per hour to ensure data continuity.

[0071] In the event of vibration sensor failure, a microphone array acoustic deep learning model is activated to identify construction noise using a historical spectrum library.

[0072] The multimodal sensing module specifically includes:

[0073] Dust monitoring unit:

[0074] A laser scattering sensor irradiates suspended particulate matter in the air with a laser beam of a specific wavelength and calculates the concentration values ​​of PM2.5 and PM10 in real time based on the intensity distribution of the scattered light from the particles.

[0075] A beta-ray sensor uses an internal radioactive source to release beta rays that penetrate a sampling filter membrane, and the mass concentration of particulate matter is accurately determined by the amount of ray attenuation.

[0076] The laser scattering sensor continuously outputs monitoring data, while the beta ray sensor automatically initiates the calibration process at regular intervals, feeding back the absolute measurement value to the processing core to dynamically correct the zero-point drift and range error of the laser sensor.

[0077] Noise monitoring unit:

[0078] The microphone array consists of high-sensitivity digital microphones arranged in a diamond topology. It uses a beamforming algorithm to spatially filter the sound field in the construction area and generate a directional pickup beam with an adjustable angle range.

[0079] The vibration sensor uses a MEMS triaxial accelerometer to collect the vibration spectrum characteristics within a predetermined wide frequency range;

[0080] The noise separation engine performs time-frequency domain matching between the sound signal and the vibration signal. When the microphone array captures a sound pressure event, it searches the synchronous vibration spectrum database. If a mechanical equipment characteristic spectrum that is time-aligned with the sound event is detected, it is determined to be valid construction noise. If the sound event has no matching vibration characteristics, it is regarded as background interference and is eliminated.

[0081] Environmental parameter unit:

[0082] The temperature and humidity sensor uses a capacitive polymer thin film probe to monitor the ambient temperature and humidity within a predetermined temperature range, providing a humidity correction coefficient for dust monitoring.

[0083] The barometer is based on a piezoresistive MEMS chip and combines temperature data to calculate air density, correcting the gas absorption effect error of the β-ray sensor.

[0084] The rain sensor detects precipitation intensity through an infrared optical grid. When a raindrop hits the monitoring window, it triggers a rainproof mode, closes the air intake valve of the laser scattering sensor to reduce the risk of wet pollution, and activates a sedimentation compensation model for dust caused by rainfall.

[0085] It should be noted that the dust monitoring unit:

[0086] Laser scattering sensor: It adopts a 785nm near-infrared laser source (avoiding interference from natural light), and retrieves the 0.3-10μm particle size spectrum through the intensity distribution of 90° side-scattered light, achieving a real-time response of 500ms and accurately capturing instantaneous dust events such as blasting and earthmoving operations.

[0087] The optical self-cleaning design features an oleophobic nanofilm coating on the lens surface, combined with a piezoelectric ceramic oscillator embedded in the cavity, which triggers infrasonic dust removal every 10 minutes to ensure light transmittance in extreme dusty environments.

[0088] Beta-ray sensor: A low-activity carbon-14 radioactive source (activity < 100 kBq) is selected, conforming to the safety standard GB / T18883-2022. After beta particles penetrate the sampling filter membrane, the attenuation is detected by a Geiger counter. The mass concentration measurement error is ≤ ±3 μg / m³. 3 ;

[0089] Intelligent filter membrane management automatically switches sampling points based on the cumulative particulate matter load (rotary filter belt contains 100 points), a single calibration consumes only 1 point, and the filter belt life is up to 3 months.

[0090] Dual-sensor collaborative mechanism:

[0091] Master-slave data fusion uses a laser sensor as the main sampler (1Hz high-frequency output) and a beta-ray sensor to output reference values ​​4 times a day. The temperature drift and nonlinear error of the laser sensor are dynamically corrected by Kalman filtering.

[0092] Seamless fault switching: When the laser fails due to mirror contamination, the beta ray sampling frequency is automatically increased to 1 time / hour to maintain monitoring continuity.

[0093] Noise monitoring unit:

[0094] Directional sensing of microphone array: Beamforming algorithm, based on MVDR (Minimum Variance Distortionless Response) algorithm to generate adaptive beams, can still extract target sound sources within a 45° cone angle under 30dB background noise; supports sound source tracking mode, with positioning error of <1.5 meters for mobile devices (such as dump trucks);

[0095] An environmental compensation mechanism dynamically corrects the sound wave attenuation model based on temperature and humidity sensor data (e.g., 3dB compensation for the increase in frequencies above 2kHz under 25℃ / 80%RH conditions).

[0096] Vibration sensor:

[0097] Wideband sensing: The MEMS accelerometer has a frequency response range of 0.5Hz-5kHz, which can capture low-frequency vibration (13Hz) of diesel engines and high-frequency impact (4kHz) of hydraulic breakers; Equipment fingerprint database construction: It has pre-stored vibration spectrum templates of 30 types of construction machinery (such as the characteristic peak of excavators being 215±10Hz), and supports online learning to add new equipment features.

[0098] Noise separation engine:

[0099] Time-frequency-space three-dimensional verification:

[0100] Time alignment: The time delay difference between the sound signal and the vibration signal is <50ms (guaranteed by Bluetooth 5.0 synchronization protocol);

[0101] Frequency domain matching: Calculate the coherence function of the acoustic vibration signal at 1 / 3 octave band > 0.65;

[0102] Spatial consistency: The distance between the microphone array positioning sound source and the vibration sensor installation location is <5 meters;

[0103] Deep learning assistance: For complex mixed noise (such as wind and rain noise + equipment noise), a CNN model is used for voiceprint separation, with a false positive rate of less than 4%.

[0104] Environmental parameter unit:

[0105] Temperature and humidity sensors: Dust monitoring compensation, establishing a humidity-scattered light intensity correction model, reducing the laser sensor error from ±28% to ±6% in a 95%RH high humidity environment; Noise monitoring compensation, temperature gradient affects sound speed, and the sound source localization algorithm is calibrated in real time based on temperature and humidity data.

[0106] Gas effect elimination of barometer: Dynamic density calculation, based on the ideal gas law, calculates the effect of air density on the penetrating power of β rays (e.g., 4.2% compensation is required in areas with an altitude of 500 meters); Storm warning linkage, when the air pressure drops by more than 5 hPa / 3 minutes, the windproof reinforcement command of the equipment is triggered;

[0107] Multi-system collaboration of rain sensor enables intelligent switching of rain protection modes:

[0108] Light rainfall (<10mm / h): Only the laser sensor intake valve is closed; the beta ray continues to operate.

[0109] Heavy rainfall (>30mm / h): Settlement compensation model activated; PM10 measured value = raw value × e -0.07t t represents the duration of rainfall;

[0110] Water resource optimization: The sprinkler system is automatically disabled during rainfall, and a "natural dust suppression efficiency report" is pushed through the cloud platform.

[0111] The environmental adaptive calibration module specifically includes:

[0112] Humidity compensation unit:

[0113] The real-time correction mechanism involves the temperature and humidity sensor collecting environmental data every second. When the relative humidity is less than or equal to a predetermined ratio, the humidity compensation algorithm is activated, and the correction coefficient is calculated in segments based on the humidity value. When the humidity value is greater than the predetermined ratio, the output is automatically switched to the β-ray sensor, and the laser scattering data is used as an auxiliary reference.

[0114] Based on the status of the raindrop sensor, if continuous precipitation is detected, the β-ray dominant mode will be retained even if the humidity value is greater than a predetermined ratio; if there is only high humidity but no precipitation, the mode will be switched back to laser dominant mode after the humidity drops back to the predetermined ratio.

[0115] Noise separation unit:

[0116] Synchronous signal acquisition: a microphone array captures sound pressure signals at a predetermined sampling rate, and a vibration sensor synchronously acquires triaxial acceleration data at the contact points of the construction machinery.

[0117] The interference filtering process includes feature extraction, which includes acoustic signals to extract typical acoustic signature features of construction equipment and vibration signals to identify the fundamental frequency harmonics of the equipment.

[0118] Correlation determination involves calculating the cross-correlation coefficient between the acoustic signal and the vibration signal within a time window. If the cross-correlation coefficient is greater than or equal to a predetermined threshold and the frequency characteristics match, it is determined to be valid construction noise.

[0119] Dynamic threshold adjustment adapts to background noise intensity at different times, based on the ambient noise background value.

[0120] Drift correction unit:

[0121] The timed calibration is triggered, and the calibration sequence is started at a set time. The β-ray sensor automatically collects particulate matter mass concentration data and records the laser scattering sensor readings at the same time.

[0122] Drift compensation mechanism, zero calibration. When the PM10 is continuously less than the established threshold during the site silent period, the baseline value of the laser scattering sensor is forced to zero; range calibration, comparing the absolute measurement value of beta ray with the laser scattering reading, generating a proportionality coefficient, dividing the proportionality coefficient range, and different ranges trigger calibration strategies;

[0123] Self-diagnosis and fault tolerance. When the membrane load of the beta ray sensor is greater than the established ratio, the calibration is automatically paused, and the solar self-cleaning module is triggered to remove dust accumulation to ensure the reliability of the calibration reference.

[0124] It should be noted that the humidity compensation unit:

[0125] Dynamic segmented compensation mechanism: Refinement of humidity threshold. The default humidity threshold is set to 80% (configurable). When RH≤60%, linear compensation is adopted: PM_real = PM_laser × (1 - 0.018RH);

[0126] When 60% < RH ≤ 80%, an exponential compensation model is enabled:

[0127] PM_real = PM_laser × e^(-0.045RH);

[0128] The compensation coefficient is obtained by fitting through high-temperature and high-humidity cycle tests in the laboratory (20℃~50℃ / 30%~95%RH);

[0129] Enhancement of beta ray dominant mode: When RH > 80%, the system takes the beta ray data as the reference value, and continuously monitors the deviation rate of the laser sensor relative to the beta ray. If the deviation rate is continuously < 5% for 2 hours, the dual-sensor weighted fusion output (beta ray weight 0.7) is automatically restored;

[0130] Precipitation linkage protection strategy, rainfall level response:

[0131] Light rain (0.1~2mm / h): Only close the intake valve of the laser sensor, and the beta ray keeps normal sampling;

[0132] Moderate to heavy rain (>2mm / h): Activate the sedimentation compensation model to perform exponential decay correction on the measured value of PM10:

[0133] PM_corrected = PM_original × e^(-0.06t); t is the rainfall duration;

[0134] Intelligent delay recovery: In high-humidity and rainless scenarios, after the humidity drops to 75%, it is delayed for 15 minutes to switch back to the laser main mode to avoid data oscillation caused by frequent switching.

[0135] Noise separation unit:

[0136] High-precision synchronous acquisition architecture: Hardware-level clock synchronization, with time synchronization via GPS module or 5G network, ensures that the time synchronization error between the microphone array and vibration sensor is <1ms; the vibration signal sampling rate is increased to 10kHz, covering the 0.5Hz~8kHz frequency band (including infrasound from impact equipment and high-frequency harmonics from metal collisions).

[0137] Deep learning: Equipment acoustic signature feature library, pre-trained CNN model to identify the acoustic signature of typical construction equipment (such as the 1.2kHz narrowband pulse of a hydraulic breaker and the 63Hz low-frequency roar of a concrete pump truck); vibration spectrum is extracted by wavelet packet decomposition to extract 16 sub-band energy features to construct the equipment vibration "fingerprint".

[0138] Dynamic correlation determination, with a short-time cross-correlation coefficient (STCC) calculation window of 0.5 seconds, and the correlation coefficient threshold dynamically adjusted according to day and night:

[0139] Daytime (7:00-22:00): Threshold = 0.7;

[0140] Nighttime (22:00-7:00): Threshold = 0.5 (adapting to lower ambient noise background);

[0141] Frequency matching requirement: The main peak frequency of the acoustic signal must fall within ±10% of the fundamental frequency of the vibration signal;

[0142] Enhanced robustness in complex scenarios, resisting wind and rain interference:

[0143] When the wind speed is >5m / s, the wind noise suppression algorithm (based on spectrum flatness detection of broadband noise) is activated; raindrop impact sound is shielded by short-time energy change detection.

[0144] Mobile sound source tracking: By combining coordinate data from the UWB positioning system, the pointing angle of the microphone array beam is dynamically adjusted to achieve noise tracking and separation of mobile devices (such as dump trucks).

[0145] Drift correction unit:

[0146] Intelligent calibration triggering strategies: Periodic calibration, which forces a 15-minute β-ray sampling every 6 hours and compares it with the laser sensor readings to generate a calibration curve; Event-driven calibration, which is triggered when the PM10 concentration suddenly exceeds 50 μg / m³. 3 At a speed of / min, an additional calibration sequence is immediately triggered to eliminate transient interference from the sensor;

[0147] Graded drift compensation mechanism: Zero-point drift suppression, automatically detecting PM10 background values ​​from 1:00 AM to 5:00 AM daily; if PM10 < 15 μg / m³ for 3 consecutive hours... 3 If the wind speed is less than 1 m / s, the laser sensor baseline is zeroed; the β-ray reading is checked simultaneously during the zeroing process to ensure that the low value is not caused by sensor failure.

[0148] Membrane load monitoring calculates the particulate matter load on the filter membrane by the slope of the decrease in the β-ray count rate. When the load rate is >85%, sampling is paused and automatic filter belt feeding is triggered. Abnormal load (such as a 50% surge in load rate within 1 hour) is determined to be a sensor failure, and the backup module is activated.

[0149] Self-cleaning enhancement: The solar brush performs optical window cleaning at 11:00 and 17:00 daily (avoiding the midday high temperature and nighttime low temperature condensation period); after cleaning, the cleaning effect is evaluated by the laser scattered light intensity recovery rate. If the recovery rate is <90%, a maintenance work order is sent.

[0150] The intelligent response and linkage module specifically includes:

[0151] Multi-parameter fusion determination mechanism:

[0152] Dynamic identification of pollution events; triggering conditions for dust exceeding standards: when the PM10 concentration exceeds the preset threshold for a continuous period of time, and the environmental parameter unit excludes precipitation interference, it is determined to be a valid dust event; triggering conditions for noise exceeding standards: when construction noise, after being filtered by the noise separation unit, exceeds the legal limit for a continuous period of time, and the vibration sensor simultaneously verifies the equipment operating status; priority of compound events: if dust and noise exceed standards at the same time, dust suppression measures are initiated first; for single events, a targeted response is triggered.

[0153] Pollution heat maps are generated based on data from multimodal sensing units, including dust source location, combining wind speed and direction sensor data and PM concentration gradient changes to invert the pollution diffusion path; noise source location, calculating the time difference of arrival of sound through microphone array beamforming to pinpoint the coordinates of equipment exceeding the standard.

[0154] Precise dust suppression implementation:

[0155] Positioning-based sprinkler control, intelligent sprinkler networking, deployment of IoT smart sprinklers, and automatic adjustment of spray angle after receiving the coordinates of the pollution source;

[0156] Dynamic water volume optimization: for light pollution, intermittent spraying from a single nozzle is activated; for heavy pollution, all nozzles within a predetermined radius are activated to work continuously, and a fog cannon truck is linked to enhance coverage; water-saving strategy: based on feedback from the raindrop sensor, if the rainfall intensity exceeds a predetermined threshold, the spraying is automatically paused to utilize natural rainfall for dust suppression.

[0157] Directional sound wave suppression and sound wave interference noise reduction: For fixed noise sources, reverse sound waves are generated through sound field modeling to form a silencing zone in the target area; dynamic tracking of moving sound sources, combined with UWB positioning module to update the phase of the reverse sound waves in real time; sound and light alarm coordination: flashing LED warnings are triggered around the equipment exceeding the standard, and directional voice prompts are played simultaneously; after a predetermined period of continuous exceeding the standard, an alarm work order is automatically pushed to the mobile phone of the responsible personnel.

[0158] Predictive linkage control:

[0159] AI-driven proactive intervention, a dust-noise correlation model, and real-time analysis of construction plan data. Input parameters include equipment start-up and shutdown plans, earthwork operation area, and short-term weather forecasts. Output actions include pre-starting the sprinkler system to wet the work surface before the predicted exceedance.

[0160] Closed-loop performance evaluation involves continuous monitoring of pollution parameter change rates after response. If the PM10 concentration does not decrease as expected within a predetermined timeframe, the spray pressure is automatically increased and additional backup spray nozzles are activated. A response performance report is generated to optimize control parameters for the next operation.

[0161] It should be noted that the multi-parameter fusion judgment mechanism is as follows:

[0162] Dynamic identification of pollution incidents, with adaptive adjustment of precise thresholds for determining dust exceeding standards:

[0163] The baseline threshold is set at 150 μg / m 3 (Complies with GB3095-2025 standard), dynamically adjusted according to the construction stage (e.g., increased to 200μg / m³ during earthwork excavation). 3 To eliminate precipitation interference, when the raindrop sensor detects precipitation >2mm / h, the PM10 exceedance judgment time is automatically extended from 5 minutes to 15 minutes to avoid false judgments caused by natural sedimentation.

[0164] For diffusion path verification, a Gaussian plume-particle diffusion hybrid model is used, based on wind speed and direction sensor data. If the concentration peak area matches the pollution diffusion trajectory downstream of the prevailing wind direction by more than 80%, it is confirmed as a valid dust event.

[0165] Noise exceeding standards verification, construction status double confirmation:

[0166] The vibration intensity detected by the vibration sensor is >0.5g (g is the acceleration due to gravity);

[0167] The camera AI recognizes the operating actions of the equipment (such as the movement of the excavator arm) and aligns them with the sound signals in time and space;

[0168] The statutory limits are dynamically switched, automatically switching between day and night standards according to local environmental regulations (e.g., 70dB during the day and 55dB at night in Beijing), and automatically lowering the limits by 5dB on holidays.

[0169] Composite event prioritization strategy:

[0170] Pollution synergy calculations, when dust and noise occur concurrently, are based on a health risk model (for every 10 μg / m³ increase in PM10...). 3 Dust suppression should be prioritized if the equivalent risk of 3dB noise is increased by 40%.

[0171] In resource conflict arbitration, if dust suppression spraying and sound wave suppression need to be activated simultaneously, priority should be given to ensuring the spraying (because water mist can simultaneously reduce noise by 6-8dB).

[0172] Spatial location of pollution sources:

[0173] Dust source inversion and localization, multi-sensor gradient analysis, and concentration difference of PM10 sensors deployed at the four corners of the construction site (e.g., 200 μg / m³ at point A) were used. 3 Point B 80μg / m 3 Based on the westerly wind speed of 2.5 m / s, the pollution source is estimated to be 15 meters east of point A.

[0174] Visual-assisted verification, linked with panoramic cameras to identify dust visibility (weighting is increased when visibility is <500 meters);

[0175] Noise source TDOA localization, microphone array optimization, 4-microphone diamond array time difference positioning accuracy reaches ±1.2 meters, fused with UWB positioning module (accuracy ±0.3 meters) to track mobile devices;

[0176] Voiceprint and fingerprint database matching; noise exceeding the standard is compared using a deep learning model and directly associated with the device ID (e.g., "excavator #07").

[0177] Precise dust suppression implementation:

[0178] Positioning-based sprinkler control, intelligent sprinkler network architecture:

[0179] Cloud-edge collaborative control: edge computing nodes analyze pollution source coordinates in real time (latency <200ms), and the central cloud platform optimizes the entire network spraying strategy; nozzle kinematics optimization: stepper motors drive the nozzles, with a horizontal rotation speed of 90° / s and a pitch adjustment accuracy of ±0.5°;

[0180] Dynamic water volume optimization algorithm:

[0181]

[0182] The water-saving AI model dynamically reduces the amount of water sprayed based on soil moisture sensor data (e.g., 20% reduction in clay soil).

[0183] Natural precipitation synergy mechanism:

[0184] When there is moderate rain (>5mm / h), turn off the sprinklers and activate the rainwater diversion system: adjust the slope of the site to direct rainwater to areas prone to dust; generate a natural dust suppression efficiency report and quantify the amount of water saved (e.g., 15 tons of water saved in a single rainfall).

[0185] Directional acoustic suppression:

[0186] Active noise reduction technology addresses fixed noise sources by generating reverse sound waves through sound field simulation, creating a 5-meter diameter silencing zone around the target equipment (noise reduction of 12dB@500Hz); the noise reduction frequency band focuses on the characteristic peaks of construction equipment (such as 200Hz for generators and 4kHz for cutting machines).

[0187] Mobile noise source tracking: UWB tag update location (10Hz frequency), dynamic phase adjustment of the sound wave transmitter; noise reduction delay for mobile sources such as dump trucks <100ms;

[0188] Multi-level response for audible and visual alarms:

[0189] Level 1 (Immediate Exceedance): Red flashing (frequency 2Hz) is triggered around the equipment, and a directional broadcast "Please reduce speed" is sent.

[0190] Level 2 (lasts 5 minutes): Push alarm work orders to the responsible person's mobile phone and upload them to the law enforcement platform simultaneously;

[0191] Level 3 (lasts 10 minutes): Automatically limits the throttle opening of the equipment (via CAN bus).

[0192] Predictive linkage control:

[0193] AI-driven proactive intervention, dust-noise correlation model, multi-dimensional fusion of input parameters:

[0194]

[0195] LSTM prediction engine:

[0196] Training data: 100,000 historical pollution event samples;

[0197] Output: When the probability of exceeding the standard is >80%, dust suppression will be pre-activated 30 minutes in advance.

[0198] Dust suppression strategy pre-loading: The working surface is pre-wetted and dust suppressant (water-based polymer film) is sprayed before earthwork excavation to reduce dust generation by 40%; Equipment noise reduction pre-adjustment: Before noise exceeds the standard, the pressure setting value of the crusher hydraulic system is automatically reduced.

[0199] Closed-loop effect evaluation:

[0200] Dynamically optimize execution parameters and evaluate dust suppression response:

[0201] If the PM10 level decreases by less than 30% within 10 minutes after spraying, the water pressure will be automatically increased by 0.2 MPa or additional spray nozzles will be opened.

[0202] If the response fails to meet the standard after 3 consecutive attempts, the system will perform a self-check (to check for nozzle blockage / sensor drift).

[0203] Noise reduction response assessment: Noise reduction <6dB after 5 minutes of sound wave suppression, switch to high-frequency focusing mode (sacrificing coverage to improve local noise reduction).

[0204] Digital Twin Optimization Engine

[0205] Build a digital dashboard for response performance:

[0206] Dust suppression rate = (PM10 before intervention - PM10 after intervention) / PM10 before intervention;

[0207] Noise reduction gain = reduction in sound pressure level in the target frequency band (dB);

[0208] Based on reinforcement learning and dynamic parameter tuning, historical response data is fed back to the AI ​​model to iteratively optimize strategies (such as increasing the frequency of pre-humidification during the high-temperature period in summer).

[0209] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A construction site dust and noise monitoring system, characterized in that, include: Multimodal sensing module: Dust monitoring unit, integrating a laser scattering sensor and a beta-ray sensor, wherein the laser scattering sensor collects PM2.5 / PM10 concentration in real time, and the beta-ray sensor is calibrated according to a preset period; Noise monitoring unit, deploying a microphone array and a vibration sensor, wherein the microphone array uses acoustic beamforming technology to directionally capture sound sources in the construction area, and the vibration sensor synchronously collects the mechanical vibration spectrum to separate background noise; The environmental parameter unit is equipped with a temperature and humidity sensor, a barometer, and a rain sensor. Environmental adaptive calibration module: Humidity compensation unit, which dynamically corrects the water mist interference error of the laser scattering sensor based on temperature and humidity data. When the humidity is greater than a predetermined ratio, it automatically switches to the β-ray sensor as the main output. The noise separation unit filters out non-construction-related traffic noise and mechanical resonance interference by comparing the vibration sensor spectrum with the microphone array acoustic signature; the drift correction unit uses the absolute measurement value of the β-ray sensor to perform automatic zero-point calibration of the laser scattering sensor at regular intervals. Intelligent response linkage module: connects the spray controller and the sound and light alarm. When the PM10 concentration exceeds the threshold for a set number of minutes or the noise exceeds the standard, it triggers the positioning spray and directional sound wave suppression.

2. The construction site dust and noise monitoring system according to claim 1, characterized in that, The multimodal sensing module specifically includes: Dust monitoring unit: A laser scattering sensor irradiates suspended particulate matter in the air with a laser beam of a specific wavelength and calculates the concentration values ​​of PM2.5 and PM10 in real time based on the intensity distribution of the scattered light from the particles. A beta-ray sensor uses an internal radioactive source to release beta rays that penetrate a sampling filter membrane, and the mass concentration of particulate matter is accurately determined by the amount of ray attenuation. The laser scattering sensor continuously outputs monitoring data, while the beta ray sensor automatically initiates the calibration process at regular intervals, feeding back the absolute measurement value to the processing core to dynamically correct the zero-point drift and range error of the laser sensor.

3. The construction site dust and noise monitoring system according to claim 2, characterized in that, The multimodal sensing module specifically includes: Noise monitoring unit: The microphone array consists of high-sensitivity digital microphones arranged in a diamond topology. It uses a beamforming algorithm to spatially filter the sound field in the construction area and generate a directional pickup beam with an adjustable angle range. The vibration sensor uses a MEMS triaxial accelerometer to collect the vibration spectrum characteristics within a predetermined wide frequency range; The noise separation engine performs time-frequency domain matching between the sound signal and the vibration signal. When the microphone array captures a sound pressure event, it searches the synchronous vibration spectrum database. If a mechanical equipment characteristic spectrum that is time-aligned with the sound event is detected, it is determined to be valid construction noise. If the sound event has no matching vibration characteristics, it is regarded as background interference and is eliminated.

4. The construction site dust and noise monitoring system according to claim 3, characterized in that, The multimodal sensing module specifically includes: Environmental parameter unit: The temperature and humidity sensor uses a capacitive polymer thin film probe to monitor the ambient temperature and humidity within a predetermined temperature range, providing a humidity correction coefficient for dust monitoring. The barometer is based on a piezoresistive MEMS chip and combines temperature data to calculate air density, correcting the gas absorption effect error of the β-ray sensor. The rain sensor detects precipitation intensity through an infrared optical grid. When a raindrop hits the monitoring window, it triggers a rainproof mode, closes the air intake valve of the laser scattering sensor to reduce the risk of wet pollution, and activates a sedimentation compensation model for dust caused by rainfall.

5. A construction site dust and noise monitoring system according to claim 4, characterized in that, The environmental adaptive calibration module specifically includes: Humidity compensation unit: The real-time correction mechanism involves the temperature and humidity sensor collecting environmental data every second. When the relative humidity is less than or equal to a predetermined ratio, the humidity compensation algorithm is activated, and the correction coefficient is calculated in segments based on the humidity value. When the humidity value is greater than the predetermined ratio, the output is automatically switched to the β-ray sensor, and the laser scattering data is used as an auxiliary reference. Based on the status of the raindrop sensor, if continuous precipitation is detected, the β-ray dominant mode will be retained even if the humidity value is greater than a predetermined ratio; if there is only high humidity but no precipitation, the mode will be switched back to laser dominant mode after the humidity drops back to the predetermined ratio.

6. A construction site dust and noise monitoring system according to claim 5, characterized in that, The environmental adaptive calibration module specifically includes: Noise separation unit: Synchronous signal acquisition: a microphone array captures sound pressure signals at a predetermined sampling rate, and a vibration sensor synchronously acquires triaxial acceleration data at the contact points of the construction machinery. The interference filtering process includes feature extraction, which includes acoustic signals to extract typical acoustic signature features of construction equipment and vibration signals to identify the fundamental frequency harmonics of the equipment. Correlation determination involves calculating the cross-correlation coefficient between the acoustic signal and the vibration signal within a time window. If the cross-correlation coefficient is greater than or equal to a predetermined threshold and the frequency characteristics match, it is determined to be valid construction noise. Dynamic threshold adjustment adapts to background noise intensity at different times, based on the ambient noise background value.

7. A construction site dust and noise monitoring system according to claim 6, characterized in that, The environmental adaptive calibration module specifically includes: Drift correction unit: The timed calibration is triggered, and the calibration sequence is started at a set time. The β-ray sensor automatically collects particulate matter mass concentration data and records the laser scattering sensor readings at the same time. Drift compensation mechanism, zero-point calibration: when PM10 is detected to be continuously lower than the predetermined threshold during the construction site's quiet period, the baseline value of the laser scattering sensor is forced to zero; range calibration: compare the absolute measurement value of β-rays with the laser scattering reading, generate a proportional coefficient, divide the proportional coefficient range, and trigger calibration strategies in different ranges. Self-diagnostic fault tolerance: When the load on the β-ray sensor membrane exceeds a predetermined ratio, calibration is automatically paused, triggering the solar self-cleaning module to remove accumulated dust and ensure the reliability of the calibration reference.

8. A construction site dust and noise monitoring system according to claim 7, characterized in that, The intelligent response and linkage module specifically includes: Multi-parameter fusion determination mechanism: Dynamic identification of pollution events; triggering conditions for dust exceeding standards: when the PM10 concentration exceeds the preset threshold for a continuous period of time, and the environmental parameter unit excludes precipitation interference, it is determined to be a valid dust event; triggering conditions for noise exceeding standards: when construction noise, after being filtered by the noise separation unit, exceeds the legal limit for a continuous period of time, and the vibration sensor simultaneously verifies the equipment operating status; priority of compound events: if dust and noise exceed standards at the same time, dust suppression measures are initiated first; for single events, a targeted response is triggered. Pollution heat maps are generated based on data from multimodal sensing units, including dust source location, combining wind speed and direction sensor data and PM concentration gradient changes to invert the pollution diffusion path; noise source location, calculating the time difference of arrival of sound through microphone array beamforming to pinpoint the coordinates of equipment exceeding the standard.

9. A construction site dust and noise monitoring system according to claim 8, characterized in that, The intelligent response and linkage module specifically includes: Precise dust suppression implementation: Positioning-based sprinkler control, intelligent sprinkler networking, deployment of IoT smart sprinklers, and automatic adjustment of spray angle after receiving the coordinates of the pollution source; Dynamic water volume optimization: for light pollution, intermittent spraying from a single nozzle is activated; for heavy pollution, all nozzles within a predetermined radius are activated to work continuously, and a fog cannon truck is linked to enhance coverage; water-saving strategy: based on feedback from the raindrop sensor, if the rainfall intensity exceeds a predetermined threshold, the spraying is automatically paused to utilize natural rainfall for dust suppression. Directional sound wave suppression and sound wave interference noise reduction: For fixed noise sources, reverse sound waves are generated through sound field modeling to form a silencing zone in the target area; dynamic tracking of moving sound sources, combined with UWB positioning module to update the phase of the reverse sound waves in real time; sound and light alarm coordination: flashing LED warnings are triggered around the equipment exceeding the standard, and directional voice prompts are played simultaneously; after a predetermined period of continuous exceeding the standard, an alarm work order is automatically pushed to the mobile phone of the responsible personnel.

10. A construction site dust and noise monitoring system according to claim 9, characterized in that, The intelligent response and linkage module specifically includes: Predictive linkage control: AI-driven proactive intervention, a dust-noise correlation model, and real-time analysis of construction plan data. Input parameters include equipment start-up and shutdown plans, earthwork operation area, and short-term weather forecasts. Output actions include pre-starting the sprinkler system to wet the work surface before the predicted exceedance. Closed-loop performance evaluation involves continuous monitoring of pollution parameter change rates after response. If the PM10 concentration does not decrease as expected within a predetermined timeframe, the spray pressure is automatically increased and additional backup spray nozzles are activated. A response performance report is generated to optimize control parameters for the next operation.

Images