Building engineering environment monitoring method and system

Abstract

The invention discloses a building engineering environment monitoring method and system, and particularly relates to the technical field of environment monitoring, and the method comprises the steps: S1, building an initial monitoring network topology structure; s2, generating composite noise source recognition features; s3, generating a node operation load degree; s4, monitoring a priority adjustment instruction; and S5, integrating the data to form a hierarchical environment state data set. According to the construction engineering environment monitoring method and system, the specific structure size of the tower crane is obtained through the laser range finder, and the initial layout of the monitoring points is realized in combination with the sound wave attenuation model, so that the traditional gridding or empirical point distribution is abandoned, and the construction engineering environment monitoring efficiency is improved. According to the method, the monitoring network has the optimized coverage capability aiming at a core dynamic noise source from the beginning of construction, the monitoring signal-to-noise ratio of a key area is improved, and in the noise identification stage, a sliding time window cross-correlation algorithm is adopted to carry out synchronous analysis on two specific noise sources of the tower crane and the pump truck.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, and in particular to a method and system for environmental monitoring of building engineering. Background Technology

[0002] The field of environmental monitoring technology includes the real-time collection and analysis of surrounding ecological indicators and pollution parameters within the work area during the construction process of building engineering. The core content of this technology involves the dynamic monitoring of key parameters such as temperature, humidity, PM2.5, noise decibels, and volatile organic compounds. By deploying sensor networks to build a data acquisition system, and combining transmission protocols to achieve remote visualization of monitoring data, its technical system covers three basic architectures: physical sensing layer, data transmission layer, and terminal processing layer.

[0003] One of the environmental monitoring methods for building engineering refers to an environmental data acquisition method based on a multi-parameter sensor deployment scheme. This method addresses three technical aspects: monitoring the particle size distribution of construction dust, analyzing the spectrum of mechanical operation noise, and detecting the concentration of volatile organic compounds in decorative materials. Specifically, dust monitoring is achieved through a laser scattering particulate matter detection unit, noise source localization is completed using a broadband sound level meter array, and volatile organic compounds are captured by an electrochemical gas sensor group. Each detection unit forms a distributed monitoring node through wireless networking technology, and the data is uploaded to the central processing platform after preliminary filtering processing by edge computing nodes.

[0004] Existing environmental monitoring technologies for construction projects have significant limitations in practical applications. Their sensor deployment schemes are usually based on a generalized grid division of the monitoring area, without being specifically optimized for core noise and vibration sources with large-scale movement characteristics, such as tower cranes, on construction sites. This results in insufficient ability of the monitoring network to capture key dynamic targets and a lack of targeted data collection. At the same time, existing technologies for noise analysis are mostly limited to isolated measurements of total sound pressure level or the spectrum of a single device, failing to effectively identify and distinguish complex noise generated by multiple large pieces of equipment working together. This makes it difficult for the monitoring system to extract specific construction activities from the noise data. For example, it is impossible to distinguish between simple vehicle transportation noise and the combined noise of concrete pumping and tower crane hoisting. Finally, the data acquisition and transmission strategy is static, with all sensors reporting data at a preset fixed frequency, lacking a dynamic adjustment mechanism. During peak periods when multiple construction activities occur simultaneously, the concurrent uploading of massive amounts of data can easily cause network congestion and exhaustion of edge computing resources, leading to the loss of critical data or processing delays. Summary of the Invention

[0005] The main objective of this invention is to provide a method and system for environmental monitoring in building engineering, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for environmental monitoring in building engineering, comprising the following steps:

[0008] S1: At the construction site, the structural dimensions and operating radius data of the tower crane are obtained by using a laser rangefinder, and the optimal geometric distance of the monitoring points is calculated by combining the sound wave attenuation model to establish the initial monitoring network topology.

[0009] S2: Call the initial monitoring network topology, synchronously collect the hoisting noise of the tower crane and the pumping noise of the concrete pump truck through the sound pressure sensor, and use the sliding time window cross-correlation algorithm to perform calculations on the two sound pressure signals, extract the energy value of the main frequency band of the same source noise, and generate composite noise source identification features;

[0010] S3: An edge unit that monitors and processes the characteristics of composite noise sources. It reads logic resource utilization, kernel temperature and data packet loss rate through the JTAG debugging interface, calculates the weighted sum and compares it with the baseline value to generate the node's operating load.

[0011] S4: Circularly compare the real-time operating load of the node with the dynamic load reference value, and generate a monitoring priority adjustment command containing sensor address and frequency parameters;

[0012] S5: Call the monitoring priority adjustment command, parse the device address and parameter fields in the monitoring priority adjustment command through the OPCUA server, adjust the reporting cycle of particulate matter monitoring in the earthwork excavation area, and integrate the data to form a hierarchical environmental status dataset.

[0013] Preferably, the initial monitoring network topology in S1 includes a set of geometric coordinate points, a sound attenuation parameter matrix, and a network topology index table;

[0014] The composite noise source identification features in S3 specifically include cross-correlation peak sequences, main frequency band energy distribution, and time-frequency feature vectors.

[0015] The real-time operating load of the nodes in S4 includes logical resource weight values, temperature correction coefficients, and network error parameters.

[0016] The monitoring priority adjustment instruction in S5 includes the device address code, frequency adjustment parameters, and instruction validity period identifier.

[0017] The hierarchical environmental status dataset in S5 includes time-stamped synchronization data streams, spatial location encoding, and quality level identifiers.

[0018] Preferably, S1-1: By acquiring the total height of the standard section and the length of the boom of the tower crane in the construction project, and collecting the horizontal distance from its rotation center to the end of the working radius, the three-dimensional envelope of the tower crane during operation is calculated by combining the two data to obtain the spatial influence range of the tower crane;

[0019] S1-2: Based on the spatial influence range of the tower crane, combined with the standard acoustic attenuation value and the equipment receiving sensitivity parameter, the spatial coordinates that meet the signal-to-noise ratio requirements in different directions are calculated in reverse, and all coordinates that meet the conditions are selected to obtain a set of monitoring candidate points.

[0020] S1-3: Call the set of candidate monitoring points, perform cluster analysis on the coordinate points in the set, select the centroid of each cluster as the deployment location, and then perform triangulation operation on these centroid points to connect adjacent points and establish the initial monitoring network topology.

[0021] Preferably, S2-1: By calling the initial monitoring network topology, two sound pressure time-domain signals are simultaneously collected at the boom rotation path of the tower crane and the feed inlet of the concrete pump truck. The two signals are then digitally sampled to generate multiple original sound pressure sequences.

[0022] S2-2: For multiple original sound pressure sequences, set a sliding time window and calculate the cross-correlation coefficient between two signals within the window. Determine the difference between the peak correlation coefficient and the preset common source threshold, extract the time period where the difference is less than the tolerance range, and obtain highly correlated noise segments.

[0023] S2-3: Based on the highly correlated noise segments, perform octave band analysis, calculate the energy proportion of different frequency bands, extract the frequency bands whose energy proportion exceeds the set energy allocation threshold, and generate composite noise source identification features.

[0024] Preferably, S3-1: By monitoring the edge computing unit responsible for processing the identification characteristics of composite noise sources, continuously reading the logic resource utilization rate of its field-programmable gate array, the core temperature of the central processing unit, and the data packet loss rate of the network interface controller, the original state parameter set of the unit is obtained.

[0025] S3-2: Call the original state parameter set of the unit, multiply the logic resource utilization rate, kernel temperature and data packet loss rate by preset weight coefficients and sum them, and subtract the sum from the baseline operating power consumption value calibrated by the unit at the factory to generate the real-time operating load of the node.

[0026] Preferably, S4-1: By calling the real-time running load of the node, its value is compared with a dynamic load reference value generated based on historical data in a loop, and the number of times the real-time running load of the node exceeds the reference value is recorded to obtain the load over-limit duration period;

[0027] S4-2: Based on the duration of the overload, determine whether the value has reached the preset adjustment trigger cycle length. When the length is reached, match an instruction sequence containing the target sensor address and frequency adjustment parameters to obtain the monitoring priority adjustment instruction.

[0028] Preferably, S5-1: By calling the monitoring priority adjustment command, the device address and parameter fields contained in the command are parsed, and a configuration command is sent to the particulate matter concentration monitoring unit in the earthwork excavation area corresponding to the address to adjust its data reporting time interval and obtain the non-critical data reporting cycle;

[0029] S5-2: For non-critical data reporting cycles, maintain the original sensor acquisition frequency associated with the composite noise source identification characteristics, integrate monitoring data from different frequency sources and label them with timestamps, arrange them in chronological order, and form a hierarchical environmental status dataset.

[0030] Preferably, a system for environmental monitoring in building engineering includes:

[0031] The system includes a monitoring network construction module, a composite noise source identification module, a node load assessment module, a monitoring strategy adjustment module, and a data hierarchical fusion module.

[0032] The monitoring network construction module is used to acquire data on the structural dimensions and operating radius of the tower crane, calculate the spatial influence range of the tower crane, and deduce the set of monitoring candidate points, thereby establishing an initial monitoring network topology and sending the topology to the composite noise source identification module.

[0033] The composite noise source identification module is used to call the initial monitoring network topology, collect multiple raw sound pressure sequences, extract highly correlated noise segments through cross-correlation operations, and generate composite noise source identification features, which are then processed by the edge computing unit.

[0034] The node load assessment module is used to monitor and process the edge computing units with composite noise source identification characteristics, obtain their original state parameter sets, generate the real-time operating load of the nodes by weighted summation and comparison with the benchmark value, and send the load to the monitoring strategy adjustment module.

[0035] The monitoring strategy adjustment module is used to receive the real-time operating load of the node, obtain the load over-limit duration period by comparing it with the dynamic load reference value, generate a monitoring priority adjustment instruction, and then send the instruction to the data hierarchical fusion module.

[0036] The data grading and fusion module is used to parse monitoring priority adjustment instructions, adjust the reporting cycle of non-critical data of specific monitoring units, and integrate monitoring data from multiple sources and heterogeneous frequencies to ultimately form a graded environmental status dataset.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. This invention uses a laser rangefinder to obtain the specific structural dimensions of a tower crane and combines this with a sound wave attenuation model to achieve the initial deployment of monitoring points. This method abandons the traditional gridded or empirical point deployment, enabling the monitoring network to have optimized coverage capabilities for core dynamic noise sources from the very beginning, thus improving the signal-to-noise ratio of key areas. In the noise identification stage, a sliding time window cross-correlation algorithm is used to simultaneously analyze two specific noise sources: tower cranes and pump trucks. This allows for the accurate extraction of composite noise characteristics generated by the collaborative operation of multiple devices from complex construction background noise, achieving a deep perception of specific construction activities. In addition, the system assesses the operating load by monitoring the logical resource utilization rate, kernel temperature, and data packet loss rate of the edge computing unit itself, and dynamically adjusts non-core monitoring tasks, such as the data reporting frequency for particulate matter monitoring in earthwork areas, based on this information. This forms a monitoring strategy with closed-loop feedback and resource self-adjustment capabilities, ensuring the ability to prioritize the processing and identification of key composite noise events when computing resources are scarce, thus avoiding data congestion. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0040] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0041] like Figure 1 As shown, a method for environmental monitoring in building engineering includes the following steps:

[0042] S1: At the construction site, the height of the standard section and the length of the boom of the tower crane are obtained by using a laser rangefinder. The horizontal distance from the center of rotation to the endpoint of the maximum working radius is measured by using a total station. Combined with the ISO9613-2 sound attenuation model, the structural dimensions are substituted into the point sound source attenuation formula to calculate the A-weighted sound pressure level of monitoring points in different directions. Spatial coordinate points whose sound pressure level attenuation values ​​meet the accuracy requirements of EN61672-1 Class 1 are selected. The monitoring area coverage network is constructed by using the Delaunay triangulation algorithm to generate the initial monitoring network topology.

[0043] S2: In the environmental monitoring system, broadband noise signals generated by tower crane hoisting operations and pumping noise signals from the hydraulic system of concrete pump trucks are simultaneously acquired by IEPE type sound pressure sensors. After the two signals are sampled by a 24-bit ADC analog-to-digital converter, the cross-correlation algorithm of the sliding time window is used to calculate the cross-correlation coefficient of the time-domain waveform. When the peak value of the correlation coefficient exceeds the preset threshold, the signal of the corresponding time period is analyzed by 1 / 3 octave band, and the frequency band with an energy proportion of more than 30% of the total energy in the range of 63Hz to 8kHz is extracted. Its energy integral value is calculated to generate composite noise source identification features containing time and frequency characteristics.

[0044] S3: During the operation of the edge computing node, the lookup table occupancy rate of the FPGA chip is read in real time through the JTAG debugging interface, the temperature sensor data of each core of the multi-core processor is obtained through the I 2C bus, the number of CRC error frames of the network interface is counted through the RMON counter, and the LUT occupancy rate is multiplied by 0.5, the highest core temperature is multiplied by 0.3, and the error frame rate is multiplied by 1.2 for weighted summation. The summation result is compared with the baseline load curve obtained under the standard test environment when the node leaves the factory by a sliding window average. When the deviation value of three consecutive sampling periods exceeds the allowable fluctuation range, the real-time running load quantification parameters of the node containing the timestamp and node ID are generated.

[0045] S4: In the central control unit, noise suppression processing is performed on the real-time operating load of the node through a sliding window filter. The processed value is compared point by point with the dynamic load reference curve established based on historical operating data. When the load value of 5 consecutive sampling points exceeds the corresponding value of the reference curve, a priority adjustment event is triggered. A data frame containing a 16-bit device address, an 8-bit instruction code, and a 24-bit frequency parameter is sent to the target sensor node through the ModbusTCP protocol to generate a monitoring priority adjustment instruction with a time validity period.

[0046] S5: At the data aggregation layer, the device address and parameter fields in the monitoring priority adjustment command are parsed by the OPCUA server. A configuration command containing the new sampling interval is sent to the PM10 laser scattering particulate matter sensor in the earthwork excavation area. At the same time, the original acquisition frequency of the vibration and noise sensors is maintained. The data streams acquired at different frequencies are time-stamped through the time series database. The Kalman filter is used to eliminate the clock drift error between different sensors. Finally, the fused data is encapsulated in the ISO8601 time format to generate a hierarchical environmental status dataset containing spatial coordinates, monitoring type and quality identifier.

[0047] To obtain the total height of standard sections and the length of the jib of a tower crane used in construction projects, a handheld laser rangefinder was used on-site to measure the QTZ80 tower crane. The dimensions of each standard section were measured to be 1.8 meters by 1.8 meters by 2.5 meters, totaling 16 standard sections. From this, the total tower height was calculated to be 40 meters. The length of the jib (outer boom) was measured to be 55 meters, and the counterweight boom length to be 15.5 meters. Simultaneously, the horizontal distance from the center of rotation to the maximum working radius endpoint was collected. These two key data points—the 40-meter tower height and the 55-meter working radius—were then combined using the following formula:

[0048]

[0049] The comprehensive influence radius of the tower crane is calculated. This radius, along with the tower height, defines the crane's operating area as a cylinder with its rotation center as the axis, a radius equal to the comprehensive influence radius, and a height of 40 meters. This cylinder represents the core three-dimensional region that affects the surrounding acoustics, vibration, and visual environment during tower crane operation. The boundary coordinate set of this region is represented by (x, y, z), where x and y satisfy certain conditions, and z ranges from 0 to 40 meters, thus obtaining the spatial influence range of the tower crane. L represents the overall influence radius of the tower crane. arm α represents the measured length of the tower crane boom. w1 Represents the dynamic correction coefficient for the operation, ω slew M represents the tower crane's angular velocity, g represents the acceleration due to gravity, and M represents the angular velocity of the tower crane's rotation. load M represents the actual mass of the current lifting load. max β represents the maximum rated lifting capacity of the tower crane at the current amplitude. v1 N represents the wind-induced sway influencing factor. sec This represents the total number of seconds for wind speed monitoring data collection. Represents the instantaneous wind speed value collected in the i-th second, formula:

[0050]

[0051] Detailed Explanation and Derivation of the Formula: This formula is used to calculate the dynamic comprehensive influence radius of a tower crane in operation. All parameters are obtained through on-site equipment monitoring or actual measurement. The specific derivation process is as follows: Obtain the measured length L of the tower crane's boom. arm The distance is 55 meters. The operational dynamic correction factor α... w1 The setting is based on statistical analysis of historical hoisting sway data of the same model of tower crane under different working conditions. Its value varies with the structural stiffness and damping characteristics of the tower crane; here, based on the characteristics of the QTZ80 model, a value of 0.2 is used. The current slewing angular velocity ω of the tower crane is monitored through the slewing mechanism controller in the tower crane's control room. slewThe rotation speed is 0.6 revolutions per minute, which translates to approximately 0.6 × 2π / 60 ≈ 0.0628 radians per second. The acceleration due to gravity (g) is taken as the standard value of 9.8 meters per second squared. The actual mass M of the current lifting load is read from the tower crane's torque limiter. load The maximum rated lifting capacity M of the tower crane is 2000 kg, and the maximum rated lifting capacity M of the tower crane is read under the current working radius of 55 meters. max The weight is 2500 kg. The wind-induced swaying factor β v1 The setting is based on the aerodynamic shape of the hoisted object (such as a precast wall panel). For plate-shaped components with a large surface area, the value is set to 0.08, which is obtained by fitting wind tunnel experimental data. Instantaneous wind speed values ​​are continuously collected for 5 seconds using an ultrasonic anemometer installed at the top of the tower arm.

[0052] That is, N sec The instantaneous wind speed value is 5. The sequence is {4.5, 4.8, 5.1, 4.9, 4.7} meters per second.

[0053] Substitute the above parameters into the formula to calculate:

[0054]

[0055] The calculated value of 55.936 meters is the comprehensive influence radius of the tower crane. This indicates that, under the current operating conditions, considering the influence of rotation and wind on the hoisted object, the actual horizontal influence range of the tower crane is about 0.936 meters longer than its static boom length of 55 meters. In subsequent steps, the definition of the spatial influence range of the tower crane will directly use this calculated radius, rather than the static boom length, so that the definition of the spatial range is more in line with the actual operating conditions.

[0056] Based on the spatial influence range of the tower crane, and combined with standard acoustic attenuation values ​​and equipment receiving sensitivity parameters, the maximum distance at which the sensor can effectively identify the sound source is calculated using the following formula: The effective monitoring distance is calculated. Then, starting from the outer boundary of the cylinder defined by the tower crane's spatial influence range, this effective monitoring distance is extended outwards. Within this three-dimensional space, a set of discrete spatial coordinate points are generated at 10-meter intervals. All coordinates that satisfy the condition of a distance less than the effective monitoring distance are selected to obtain a set of candidate monitoring points. L represents the effective monitoring distance. W L represents the reference sound power level of the core noise source of the tower crane. bg C represents the background noise level of the site environment. snr This represents the minimum signal-to-noise ratio required for the sensor to effectively identify data. Represents the overall atmospheric absorption factor, H rel Represents relative humidity of the air. δ represents the overall influence radius of the tower crane calculated in the previous steps. g Represents the overall attenuation correction factor for the ground surface and obstacles;

[0057] formula: Detailed Explanation and Derivation of the Formula: This formula is used to calculate the effective monitoring distance of the sensor, taking into account environmental background noise, atmospheric absorption, signal-to-noise ratio requirements, and ground influence. Its parameters are obtained through on-site monitoring or based on standard settings. The specific derivation process is as follows: Obtain the reference sound power level L at 1 meter for the core noise source of the tower crane, namely the motor and gearbox. W The noise level was 115 decibels, which is the factory calibration value of the equipment. Before construction, an AWA6228+ multi-functional sound level meter was used to measure the noise level at the boundary of the monitoring area during a quiet period (23:00 at night) without large machinery operations, and the average value was taken to obtain the ambient background noise level L. bg It is 52 decibels;

[0058] The minimum signal-to-noise ratio C required for effective sensor identification snr According to the sensor's technical manual, to ensure the clarity of signal recognition, its value is set to 12 decibels, taking into account the atmospheric absorption factor. This value is an empirical value set based on the attenuation rate of the main frequency components of tower crane noise (concentrated between 500Hz and 1kHz) under standard atmospheric conditions. Here, it is set to 0.0004. The current relative humidity H is obtained by monitoring the air through a small on-site weather station. rel The value is 65%, which is substituted into the tower crane's comprehensive influence radius in the preceding steps. The calculated value is 55.936 meters, and the comprehensive attenuation correction coefficient δ for the ground surface and obstacles is... g Based on the site survey results, the construction site surface is mainly hardened cement flooring and compacted soil, which is highly reflective. Additionally, there are a few low-rise temporary buildings nearby. After comprehensive evaluation, a correction factor of 1.03 is set. The above parameters are then substituted into the formula for calculation.

[0059]

[0060] The presence of a negative value within the square root indicates that at a distance of 345.82 meters from the center of the tower crane, the sound pressure level has attenuated to the level of background noise plus the signal-to-noise ratio. This distance is less than the comprehensive influence radius of the tower crane. This is physically impossible, indicating a problem with the calculation. The formula needs to be re-examined. therefore And strength Distance decay L p =L W -20log 10 (r), so therefore L herep It is the smallest signal that the sensor can receive, i.e., L. bg +C snr The corrected formula is:

[0061]

[0062] The calculated result of 304.854 meters represents the effective monitoring distance. This indicates that, under the current conditions, the sensor can be effectively deployed at a maximum distance of approximately 305 meters outwards from the boundary of the tower crane's dynamic influence range. This distance will serve as the spatial basis for generating the subsequent set of candidate monitoring points. All candidate points must be located within a circle centered on the tower crane's center with a radius of [missing information]. That is, it is inside the circle with a radius of (55.936 + 304.854) = 360.79 meters, and outside the circle with a radius of 55.936 meters.

[0063] The system calls upon a set of candidate monitoring points and performs cluster analysis on thousands of coordinate points within the set. Specifically, it presets the number of monitoring nodes to be deployed, K, to be 12. Using these coordinate points as input, it randomly selects 12 points as initial centroids, calculates the Euclidean distance from each of the remaining points to these 12 centroids, and assigns each point to the nearest centroid, forming 12 initial clusters. Then, it recalculates the mean coordinates of all points within each cluster to obtain the new centroid positions. Iterates the distance calculation and centroid update steps until the position change of all centroids is less than 0.5 meters after two consecutive iterations. At this point, the 12 centroids are the final sensor deployment locations. Then, it performs triangulation on these 12 centroids, that is, it connects adjacent points to form triangles based on the principle that no three points are collinear, and requires that the circumcircle of all triangles does not contain any other centroids, thus forming a non-overlapping triangular mesh covering the entire monitoring area and establishing the initial monitoring network topology.

[0064] The initial monitoring network topology is invoked, and sensors are deployed in the area below the swing path of the tower crane boom and at the fixed feed inlet of the concrete pump truck as identified in the network. Specifically, three nodes are selected below the boom and two nodes are selected around the pump truck. IEPE type sound pressure sensors in these nodes are used to synchronously acquire two independent sound pressure time-domain signals at a sampling frequency of 48kHz. For example, within a certain second, the tower crane sensor acquires a sequence containing 48,000 voltage floating-point numbers, representing the continuous fluctuation of the sound pressure within that second. At the same time, the pump truck sensor also generates its corresponding sequence of 48,000 voltage values. These continuous voltage analog signals are quantized through a 24-bit analog-to-digital converter (ADC) and converted into discrete digital signal streams. Each signal stream consists of a timestamp and the corresponding quantization amplitude, generating multiple original sound pressure sequences.

[0065] For multiple raw sound pressure sequences, a time window with a length of 2048 sampling points is set, and the two signal sequences are synchronously slid over with a step size of 1024 sampling points. The cross-correlation coefficient R(τ) between the two signals within each time window is calculated. Specifically, one signal is fixed, and the other signal is shifted by τ sampling points on the time axis. Then, the sum of the products of corresponding points in the two sequences is calculated, and the peak value that makes R(τ) reach its maximum value is found. Then, this peak value is compared with a preset homogeneity threshold. The homogeneity threshold is set to 0.85 based on the correlation test results of noise of the same type of equipment in the laboratory. If the calculated peak value of the cross-correlation coefficient, for example, 0.91, has a difference of 0.06 from 0.85, which is less than the set tolerance range of 0.1, then it is determined that the two noise sources are strongly correlated within the time period corresponding to the 2048 sampling points. The signal data of this time period is extracted to obtain the highly correlated noise segment.

[0066] Based on the highly correlated noise segments, a 1 / 3 octave band analysis is performed on each extracted segment data. This involves passing the signal through a set of bandpass filters with center frequencies conforming to the ISO266 standard, calculating the sound pressure level energy in each 1 / 3 octave band within the range of 20Hz to 16kHz, and then calculating the percentage of energy in each band relative to the total energy of the segment. For example, in a segment with a total energy of 10 Pascals per second, the energy in the band with a center frequency of 125Hz is found to be 3.5 Pascals per second, accounting for 35% of the total energy. This percentage is then compared with the set energy allocation threshold of 30%. Since 35% exceeds 30%, the 125Hz band is identified as the dominant band of the segment. Finally, the timestamp, cross-correlation peak, dominant band center frequency, and dominant band energy value of the segment are integrated into a data vector to generate composite noise source identification features.

[0067] The edge computing unit responsible for monitoring and processing the characteristics of composite noise sources is an embedded device deployed at the construction site. Through the onboard JTAG debugging interface, it continuously reads the current number and total number of LUTs inside its core FPGA chip at a frequency of 10 times per second to calculate the logic resource utilization rate. For example, if it reads that 60,000 LUTs are in use and the total number of LUTs is 80,000, the utilization rate is 75%. At the same time, it accesses the temperature sensor built into the system-on-chip (SoC) through the I2C bus to obtain the real-time junction temperature readings of its four ARM cores, which are 65 degrees Celsius, 62 degrees Celsius, 64 degrees Celsius, and 63 degrees Celsius, respectively. It also counts the number of CRC check error data frames received by its gigabit Ethernet physical interface in the past second through the device's internal RMON counter. For example, if 5 error frames are counted, the original state parameter set of the unit is obtained.

[0068] The original state parameter set of the calling unit is then used to calculate the weighted sum using the formula:

[0069]

[0070] The node dynamic load index is calculated by subtracting the index value from the baseline operating power consumption value of 1.2, which is calibrated at the factory under full load in a standard environment of 25 degrees Celsius. The difference is the deviation between the node dynamic load index and the baseline value, and the real-time operating load of the node is generated. Represents the dynamic load index of nodes. U represents the basic weight of logical resource utilization. LUT This represents the utilization rate of logic resources in a field-programmable gate array (FPGA). The base weights representing core temperature T represents the highest real-time core temperature in a multi-core processor. amb T represents the reference ambient temperature. spec The maximum permissible operating temperature defined by the chip specification is represented by γ3, the temperature change sensitivity coefficient is represented by t, the current time point is represented by Δt, and the temperature sampling time interval is represented by M. pkt The total number of periods representing network data packet statistics. This represents the network state weights in the j-th statistical period. This represents the number of CRC check error packets detected within the j-th statistical period. This represents the total number of data packets received during the j-th statistical period;

[0071] formula:

[0072]

[0073] Detailed Explanation and Derivation of the Formula: This formula is used to calculate an index that comprehensively reflects the dynamic load of edge computing units. Its parameters are obtained through real-time monitoring and preset configuration. The specific derivation process is as follows: Basic weight of logical resource utilization. Based on long-term operational data analysis, its contribution to the overall load was determined and set to 0.45. The FPGA logic resource utilization rate U was obtained from the unit's original state parameter set. LUT The base weight for core temperature is 0.75. Also based on long-term data analysis, its impact on equipment reliability is significant, and it is set to 0.35. The highest real-time core temperature in the multi-core processor is obtained from the unit's original state parameter set. At the current time point t, the temperature is 65 degrees Celsius. Simultaneously, the temperature at the previous sampling time point t-1 is obtained as 64.5 degrees Celsius, with reference to the ambient temperature T. amb The standard ambient temperature for device calibration is set to 25 degrees Celsius, and the maximum permissible operating temperature T is defined by the chip specifications. specAccording to its datasheet, the temperature is 100 degrees Celsius, the temperature change sensitivity coefficient γ3 is an empirical value used to amplify the impact of drastic temperature fluctuations, and is set to 0.15. The temperature sampling time interval Δt is set to 1 second, and the total number of network data packet statistics periods M is... pkt The value is set to 3, meaning the data from the most recent three statistical periods is averaged, which determines the network state weights. This setting can be dynamically adjusted based on network service quality requirements. Here, it is set to 1.2 for all three periods. The number of CRC check error packets in the most recent three statistical periods is obtained through the network interface controller. The numbers are {5, 8, 6}, representing the total number of data packets received within the corresponding period. Given {150230, 149860, 151050}, substitute these parameters into the formula to calculate:

[0074]

[0075]

[0076] The calculated result of 0.60375 is the value of the node dynamic load index. This index integrates information from four dimensions: computing resources, physical temperature, temperature change rate, and network status. Compared with a simple static weighted summation, it can more sensitively reflect the instantaneous operating pressure of the node. Subsequently, this index value of 0.60375 is compared with the baseline power consumption value of 1.2. The difference of -0.59625 is the real-time operating load of the node. A negative value indicates that the current node's overall load is lower than the factory-calibrated full-load baseline.

[0077] The system calls the node's real-time operating load level and compares its value, for example, 0.3349, with a dynamically generated load reference value. This dynamic load reference value is based on the arithmetic mean of the node's real-time operating load level at the same hour (e.g., 10:00 AM) over the past 7 calendar days. Assuming the average load level at 10:00 AM over the past 7 days is 0.2500, and the current value of 0.3349 exceeds the reference value of 0.2500, the system increments the value of an internal counter. This comparison operation is performed every 10 seconds. If the node's real-time operating load level still exceeds the updated dynamic reference value in the next 10 seconds, the counter continues to accumulate; otherwise, if it is lower than the reference value, the counter is reset to zero. By continuously recording the current value of this counter, the duration of the load exceeding the limit is obtained.

[0078] Based on the duration of the overload, the system determines whether the value has reached the preset adjustment trigger cycle length. This length is set to 30 according to the equipment's stable operation requirements. That is, if the counter value continuously accumulates to 30 (representing 300 consecutive seconds, or 5 minutes, with the load continuously exceeding the limit), the system determines that intervention is required. At this time, the system will match a low-priority monitoring task associated with the current overload node according to preset rules, such as a particulate matter monitoring task in the earthwork excavation area, and generate an instruction sequence containing the Modbus device address of the particulate matter sensor (e.g., 0x0A), the function code (e.g., 0x06 for writing to the register), and the parameters for setting a new sampling frequency (e.g., adjusting the corresponding register value from once every 10 seconds to once every 60 seconds), thus obtaining the monitoring priority adjustment instruction.

[0079] The monitoring priority adjustment command is invoked. The OPCUA server in the data aggregation layer receives and parses the device address 0x0A and parameter fields contained in the command, and identifies the target device as a PM10 laser scattering particulate matter concentration monitoring unit located in the earthwork excavation area. The server then sends a configuration command to the controller of the unit. This command follows the Modbus RTU protocol and writes the frequency parameter in the command into a specific register address of the sensor control sampling cycle, thereby changing the time interval for data acquisition and reporting from the original 10 seconds to 60 seconds. This reduces the data processing volume of this non-critical task, thereby adjusting its data reporting time interval and obtaining a non-critical data reporting cycle.

[0080] In response to the adjustment of the reporting cycle for non-critical data, the reporting frequency of some data (such as PM10 particulate matter) in the system has been changed to once every 60 seconds. At the same time, the original 1-second sampling frequency of the noise sensor associated with the composite noise source identification characteristics is maintained. When the data integration module receives data streams of these two different frequencies, it marks each data with a UTC timestamp accurate to milliseconds and inserts them into a data queue arranged in chronological order. For particulate matter data, the values ​​of the previous sampling point are used to fill the 59 second-level time points between two 60-second sampling points, and are marked as "forward filling value" in the data quality field. Finally, the noise data, the original particulate matter data, and the filled particulate matter data are integrated together to form a hierarchical environmental status dataset.

[0081] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for environmental monitoring in building engineering, characterized in that, The method includes the following steps: S1: At the construction site, the structural dimensions and operating radius data of the tower crane are obtained by using a laser rangefinder, and the optimal geometric distance of the monitoring points is calculated by combining the sound wave attenuation model to establish the initial monitoring network topology. S2: Call the initial monitoring network topology, synchronously collect the hoisting noise of the tower crane and the pumping noise of the concrete pump truck through the sound pressure sensor, and use the sliding time window cross-correlation algorithm to perform calculations on the two sound pressure signals, extract the energy value of the main frequency band of the same source noise, and generate composite noise source identification features; S3: An edge unit that monitors and processes the characteristics of composite noise sources. It reads logic resource utilization, kernel temperature and data packet loss rate through the JTAG debugging interface, calculates the weighted sum and compares it with the baseline value to generate the node's operating load. S4: Circularly compare the real-time operating load of the node with the dynamic load reference value, and generate a monitoring priority adjustment command containing sensor address and frequency parameters; S5: Call the monitoring priority adjustment command, parse the device address and parameter fields in the monitoring priority adjustment command through the OPCUA server, adjust the reporting cycle of particulate matter monitoring in the earthwork excavation area, and integrate the data to form a hierarchical environmental status dataset.

2. The method for environmental monitoring of building engineering according to claim 1, characterized in that: The initial monitoring network topology in S1 includes a set of geometric coordinate points, a sound attenuation parameter matrix, and a network topology index table. The composite noise source identification features in S3 specifically include cross-correlation peak sequences, main frequency band energy distribution, and time-frequency feature vectors. The real-time operating load of the nodes in S4 includes logical resource weight values, temperature correction coefficients, and network error parameters. The monitoring priority adjustment instruction in S5 includes the device address code, frequency adjustment parameters, and instruction validity period identifier. The hierarchical environmental status dataset in S5 includes time-stamped synchronization data streams, spatial location encoding, and quality level identifiers.

3. The method for environmental monitoring of building engineering according to claim 1, characterized in that: S1-1: By acquiring the total height of the standard section and the length of the boom of the tower crane in the construction project, and collecting the horizontal distance from its rotation center to the end of the working radius, the three-dimensional envelope of the tower crane during operation is calculated by combining the two data to obtain the spatial influence range of the tower crane. S1-2: Based on the spatial influence range of the tower crane, combined with the standard acoustic attenuation value and the equipment receiving sensitivity parameter, the spatial coordinates that meet the signal-to-noise ratio requirements in different directions are calculated in reverse, and all coordinates that meet the conditions are selected to obtain a set of monitoring candidate points. S1-3: Call the set of candidate monitoring points, perform cluster analysis on the coordinate points in the set, select the centroid of each cluster as the deployment location, and then perform triangulation operation on these centroid points to connect adjacent points and establish the initial monitoring network topology.

4. The method for environmental monitoring of building engineering according to claim 1, characterized in that: S2-1: By calling the initial monitoring network topology, two sound pressure time-domain signals are simultaneously collected at the boom rotation path of the tower crane and the feed inlet of the concrete pump truck. The two signals are digitally sampled to generate multiple original sound pressure sequences. S2-2: For multiple original sound pressure sequences, set a sliding time window and calculate the cross-correlation coefficient between two signals within the window. Determine the difference between the peak correlation coefficient and the preset common source threshold, extract the time period where the difference is less than the tolerance range, and obtain highly correlated noise segments. S2-3: Based on the highly correlated noise segments, perform octave band analysis, calculate the energy proportion of different frequency bands, extract the frequency bands whose energy proportion exceeds the set energy allocation threshold, and generate composite noise source identification features.

5. The method for environmental monitoring of building engineering according to claim 1, characterized in that: S3-1: By monitoring the edge computing unit responsible for processing the identification characteristics of composite noise sources, continuously read the logic resource utilization rate of its field-programmable gate array, the core temperature of the central processing unit, and the data packet loss rate of the network interface controller, the original state parameter set of the unit is obtained. S3-2: Call the original state parameter set of the unit, multiply the logic resource utilization rate, kernel temperature and data packet loss rate by preset weight coefficients and sum them, and subtract the sum from the baseline operating power consumption value calibrated by the unit at the factory to generate the real-time operating load of the node.

6. The method for environmental monitoring of building engineering according to claim 1, characterized in that: S4-1: By calling the real-time running load of the node, its value is compared with a dynamic load reference value generated based on historical data in a loop, and the number of times the real-time running load of the node exceeds the reference value is recorded to obtain the duration of the load exceeding the limit; S4-2: Based on the duration of the overload, determine whether the value has reached the preset adjustment trigger cycle length. When the length is reached, match an instruction sequence containing the target sensor address and frequency adjustment parameters to obtain the monitoring priority adjustment instruction.

7. The method for environmental monitoring of building engineering according to claim 1, characterized in that: S5-1: By calling the monitoring priority adjustment command, the device address and parameter fields contained in the command are parsed, and a configuration command is sent to the particulate matter concentration monitoring unit in the earthwork excavation area corresponding to the address to adjust its data reporting time interval and obtain the non-critical data reporting cycle. S5-2: For non-critical data reporting cycles, maintain the original sensor acquisition frequency associated with the composite noise source identification characteristics, integrate monitoring data from different frequency sources and label them with timestamps, arrange them in chronological order, and form a hierarchical environmental status dataset.

8. A system applied to the environmental monitoring method for building engineering according to any one of claims 1-7, characterized in that, The system includes: The system includes a monitoring network construction module, a composite noise source identification module, a node load assessment module, a monitoring strategy adjustment module, and a data hierarchical fusion module. The monitoring network construction module is used to acquire data on the structural dimensions and operating radius of the tower crane, calculate the spatial influence range of the tower crane, and deduce the set of monitoring candidate points, thereby establishing an initial monitoring network topology and sending the topology to the composite noise source identification module. The composite noise source identification module is used to call the initial monitoring network topology, collect multiple raw sound pressure sequences, extract highly correlated noise segments through cross-correlation operations, and generate composite noise source identification features, which are then processed by the edge computing unit. The node load assessment module is used to monitor and process the edge computing units with composite noise source identification characteristics, obtain their original state parameter sets, generate the real-time operating load of the nodes by weighted summation and comparison with the benchmark value, and send the load to the monitoring strategy adjustment module. The monitoring strategy adjustment module is used to receive the real-time operating load of the node, obtain the load over-limit duration period by comparing it with the dynamic load reference value, generate a monitoring priority adjustment instruction, and then send the instruction to the data hierarchical fusion module. The data grading and fusion module is used to parse monitoring priority adjustment instructions, adjust the reporting cycle of non-critical data of specific monitoring units, and integrate monitoring data from multiple sources and heterogeneous frequencies to ultimately form a graded environmental status dataset.

Images