Building construction carbon emission dynamic monitoring system and method

Abstract

The invention relates to the technical field of building carbon emission monitoring, and discloses a building construction carbon emission dynamic monitoring system and method. The method comprises the following steps: acquiring carbon emission activity data in a construction process, and constructing a carbon emission dynamic monitoring database corresponding to a current construction stage; configuring a dynamic optimization target based on the database, wherein the dynamic optimization target comprises optimization of energy consumption, a material turnover rate, equipment operation efficiency and a transportation path; target classification identifiers are determined and divided into a main optimization target, an auxiliary optimization target and a benchmark maintaining target; key monitoring parameters are selected according to the main optimization target and the auxiliary optimization target, and dynamic optimization target influence analysis is carried out; and establishing a parameter optimization constraint according to an analysis result, carrying out parameter dynamic optimization, and updating a carbon emission dynamic monitoring strategy according to an optimization result. According to the method, construction carbon emission data can be dynamically captured, and the accuracy and management efficiency of carbon emission monitoring are improved through multi-dimensional target collaborative optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building carbon emission monitoring, in particular to a building construction carbon emission dynamic monitoring system and method. BACKGROUND

[0002] Under the background of increasingly severe global climate change, the building industry, as a key field of energy consumption and carbon emission, has attracted widespread attention for its carbon emission reduction work. The construction phase, as an important part of the building life cycle, involves a large amount of energy consumption, material transportation and use, mechanical equipment operation and other activities, with complex carbon emission sources and significant dynamic changes. Currently, building construction carbon emission monitoring mostly adopts a static accounting method, which estimates the carbon emission of the entire construction process based on fixed calculation models and historical data. This method often ignores the dynamic characteristics of the construction phase, such as energy demand fluctuations in different construction links, time efficiency differences in material supply, real-time changes in equipment working conditions, and dynamic adjustments in transportation routes. The static monitoring method is difficult to capture the impact of these dynamic factors on carbon emission, resulting in a large deviation between the monitoring results and the actual situation. The existing monitoring system lacks systematic analysis of carbon emission influencing factors, making it difficult to determine the contribution of different construction links, equipment types, and material types to carbon emission. This makes it difficult for construction units to develop targeted emission reduction strategies, and the implementation effect of emission reduction measures is not good. In addition, the data collection of traditional monitoring methods relies on manual recording, which is not only inefficient, but also prone to data omission or errors, further affecting the accuracy and reliability of the monitoring data. With the increasing demand for green construction and low-carbon development in the building industry, the existing static and extensive carbon emission monitoring method cannot meet the needs of fine management. How to achieve dynamic and accurate monitoring of building construction carbon emission and develop effective optimization strategies based on monitoring results has become an important issue facing the building industry. SUMMARY

[0003] The purpose of the present application is to provide a building construction carbon emission dynamic monitoring system and method to solve the problems raised in the background.

[0004] To achieve the above purpose, the present application provides a building construction carbon emission dynamic monitoring method, which comprises: Obtaining carbon emission activity data of the building construction process and establishing a carbon emission dynamic monitoring database mapped with the current construction phase; Configuring a dynamic optimization target according to the carbon emission dynamic monitoring database, the dynamic optimization target including energy consumption optimization, material turnover rate optimization, equipment operation efficiency optimization, and transportation route optimization; determining a target classification identifier in the dynamic optimization target, the target classification identifier including a main optimization target, an auxiliary optimization target, and a benchmark maintenance target; after selecting the key monitoring parameters using the main optimization target and the auxiliary optimization target, performing dynamic optimization target influence analysis on the key monitoring parameters; after establishing parameter optimization constraints according to the dynamic optimization target influence analysis results, performing parameter dynamic optimization, and updating the carbon emission dynamic monitoring strategy using the parameter dynamic optimization results.

[0005] Preferably, after selecting the key monitoring parameters using the main optimization target and the auxiliary optimization target, performing dynamic optimization target influence analysis on the key monitoring parameters includes: obtaining a set of adjustable parameters of the construction process; quantitatively mapping the influence degree of the set of adjustable parameters on the main optimization target and the auxiliary optimization target; constructing a target influence matrix of all optimization targets in the dynamic optimization target, the target influence matrix representing the mutual relationship between different optimization targets, the mutual relationship including a positive synergistic relationship and a negative constraint relationship; calculating the sensitivity coefficient of the set of adjustable parameters according to the influence degree quantitative mapping and the target influence matrix; establishing dynamic optimization target influence analysis results according to the sensitivity coefficient calculation results.

[0006] Preferably, the parameter dynamic optimization includes: after establishing the control threshold interval of the key monitoring parameters, creating an initial solution set based on the current parameter set; after performing solution fitness evaluation in the initial solution set, establishing an optimization direction and an optimization step size through parameter optimization constraints and fitness evaluation results; performing initial solution set iterative updating using the optimization direction and the optimization step size; outputting a parameter dynamic optimization scheme according to the iterative updating results.

[0007] Preferably, the initial solution set iterative updating using the optimization direction and the optimization step size includes: establishing an iteration trajectory for each solution, and identifying the iteration trajectory through the solution fitness value of each iteration; configuring an iteration evaluation window, identifying the update state of the iteration trajectory in the iteration evaluation window, generating a trajectory classification identifier, the trajectory classification identifier including a good solution trajectory classification, an exploration trajectory classification, and a poor solution trajectory classification; performing search self-optimization management of the iterative updating according to the trajectory classification identifier.

[0008] Preferably, the search self-optimization management iteratively updated according to the trajectory classification identifier comprises: A local prediction model is configured according to the trajectory classification of the optimal solution, and an improved trend prediction is performed using the local prediction model to generate a first reference optimization direction; A penalty identification layer is configured according to the trajectory classification of the suboptimal solution, and an error improvement direction is identified using the penalty identification layer to establish a window improvement taboo; The solution fine-tuning iteration update within the optimal solution trajectory classification is performed using the first reference optimization direction and the window improvement taboo, the hybrid exploration iteration update of the exploration trajectory classification is performed using the first reference optimization direction and the window improvement taboo, and the iteration update of the solution within the suboptimal solution trajectory classification is performed using the random disturbance factor.

[0009] Preferably, the execution parameter dynamic optimization further comprises: A carbon emission intensity evaluation function of the monitoring parameter is established; A balance adaptation function is established according to the carbon emission intensity evaluation function and the construction efficiency evaluation function; The optimization scheme screening of the parameter dynamic optimization is performed based on the balance adaptation function, and the optimization scheme screening result is output as the parameter dynamic optimization result.

[0010] Preferably, the method further comprises: The construction stable stage and the construction change stage are divided based on a spatiotemporal feature segmentation method; The monitoring chain tree structure of the device node and the process edge is constructed in the construction stable stage; The hierarchical optimization strategy is established according to the monitoring chain tree structure.

[0011] Preferably, the spatiotemporal feature segmentation method for dividing the construction stable stage and the construction change stage comprises: Time series carbon emission intensity data of different construction areas are obtained; The first-order difference sequence of the time series carbon emission intensity data is calculated, and the feature mutation time is screened according to the first-order difference sequence; The feature mutation time is taken as a segmentation point to divide the construction stable stage and the construction change stage.

[0012] Preferably, the method further comprises: The dynamic monitoring data stream is injected in the construction change stage; The real-time calibration of the key monitoring parameter is performed based on the cooperative analysis mechanism of the static monitoring model and the dynamic monitoring data stream; The node weight of the monitoring chain tree structure is updated according to the real-time calibration result.

[0013] Preferably, the present application also includes a construction carbon emission dynamic monitoring system for implementing the construction carbon emission dynamic monitoring method as described above, the system comprising: a data acquisition module for acquiring carbon emission activity data of the construction process and establishing a carbon emission dynamic monitoring database mapped with the current construction stage; a target configuration module for configuring dynamic optimization targets according to the carbon emission dynamic monitoring database, the dynamic optimization targets including energy consumption optimization, material turnover rate optimization, equipment operation efficiency optimization, and transportation path optimization; a classification identification module for determining target classification identifications in the dynamic optimization targets, the target classification identifications including main optimization targets, auxiliary optimization targets, and benchmark maintenance targets; an influence analysis module for performing dynamic optimization target influence analysis of key monitoring parameters after the main optimization targets and auxiliary optimization targets are selected; a dynamic optimization module for performing parameter dynamic optimization after parameter optimization constraints are established according to the dynamic optimization target influence analysis results, and updating carbon emission dynamic monitoring strategies using the parameter dynamic optimization results.

[0014] Compared with the prior art, the present application has the following beneficial effects: By establishing a carbon emission dynamic monitoring database mapped with the current construction stage, real-time and accurate capture of carbon emission activity data in the construction process is achieved, breaking the limitations of data lag and one-sidedness in the traditional static monitoring method. The dynamic association of the database with the construction stage enables carbon emission data of different construction links to be collected and analyzed in a targeted manner, providing comprehensive and practical basic data for subsequent optimization work. In the configuration of dynamic optimization targets, multiple dimensions such as energy consumption, material turnover rate, equipment operation efficiency, and transportation path are covered, forming a multi-dimensional optimization system. This multi-target optimization setting avoids the problem of increased carbon emissions in other links caused by single-target optimization, ensuring the systematicness and integrity of emission reduction work. Through the coordinated optimization of these targets, emission reduction potential can be fully tapped from multiple key nodes in the construction process. The determination of target classification identifications divides the dynamic optimization targets into main optimization targets, auxiliary optimization targets, and benchmark maintenance targets, clarifying the priority and role of different targets in the optimization process. The main optimization targets indicate the core direction of emission reduction work, the auxiliary optimization targets complement and cooperate with the main optimization targets, and the benchmark maintenance targets ensure the stability of basic performance in the optimization process. The three work together to make the optimization work both focused and comprehensive, improving the scientificity and operability of the optimization strategy.

[0015] After selecting the key monitoring parameters using the main optimization target and the auxiliary optimization target, dynamic optimization target influence analysis is performed, which can clearly grasp the influence degree and mechanism of each key parameter on carbon emissions. This in-depth analysis process avoids blind optimization, making the subsequent parameter optimization more targeted. By establishing parameter optimization constraints and performing dynamic parameter optimization, the optimal parameter combination can be found to effectively control carbon emissions while meeting the basic requirements of construction. The parameter dynamic optimization results are used to update the carbon emission dynamic monitoring strategy, forming a closed-loop mechanism of "monitoring-analysis-optimization-update". This mechanism enables the monitoring strategy to be adjusted in real time according to the changes in the construction stage and optimization results, ensuring that the monitoring work always keeps pace with the actual situation of construction and the demand for emission reduction. With the continuous updating of the monitoring strategy, the adaptability and effectiveness of the entire monitoring system are continuously improved, providing long-term support for the low-carbonization of building construction and promoting the development of the construction process towards more efficient and environmentally friendly directions. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 a timing diagram of the building construction carbon emission dynamic monitoring method described in the present application; Figure 2 a flowchart of dynamic optimization target influence analysis of key monitoring parameters; Figure 3 a flowchart of parameter dynamic optimization execution; Figure 4 a flowchart of search self-optimization management; Figure 5 a flowchart of construction stage division and layered optimization strategy. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Please refer to Figure 1 The present application provides a building construction carbon emission dynamic monitoring method, which comprises: The carbon emission activity data of energy consumption, material transportation, and equipment operation during the construction process are collected in real time by the Internet of Things sensing devices and the construction management system to build a dynamic monitoring database that is updated synchronously with the construction progress. A multi-objective optimization system is established based on database analysis, including energy consumption optimization, material turnover rate optimization, equipment operation efficiency optimization, and transportation path optimization. The optimization objectives are classified and identified, distinguishing between main optimization objectives, auxiliary optimization objectives, and baseline maintenance objectives. Dynamic optimization target influence analysis is conducted on key monitoring parameters to establish parameter optimization constraint conditions. Intelligent optimization algorithms are used for dynamic parameter optimization, and the carbon emission monitoring strategy is updated in real time based on the optimization results to form a closed-loop optimization system.

[0019] Example 1: refer to Figure 2 The adjustable parameter set of the construction process is obtained through a distributed sensing network, covering the main emission sources on the construction site. The equipment operation parameters include tower crane lifting speed, concrete pumping frequency, and air compressor start-stop cycle. The material management parameters involve reinforcement cutting scrap ratio, formwork turnover interval time, and prefabricated component approach batch. The energy distribution parameters include different construction area electricity allocation, fuel equipment dispatching priority, and temporary heating temperature threshold. The data collection module is synchronized in real time with the engineering progress planning system to ensure that the parameter set reflects the characteristics of the current construction stage. The parameter set is stored in the dynamic monitoring database, and the index relationship is established according to the equipment type, process number, and time stamp.

[0020] The optimization target quantitative analysis process adopts a two-stage mapping mechanism. For the main optimization objectives, such as equipment operation efficiency optimization, a target parameter correlation model is established to determine the core influence factor set. Taking a tower crane as an example, the operating parameters include lifting speed, standby time, and the number of simultaneous operation equipment. The target function is used to calculate the weight coefficient to evaluate the contribution of each parameter to shortening the equipment operation cycle. The auxiliary optimization objectives use relative weight allocation, such as the formwork reuse parameter in material turnover rate optimization, to analyze their indirect influence on energy consumption optimization. The quantitative process combines construction progress data for discretization processing to convert the target influence degree into a five-level scale value, forming a parameter target influence scale.

[0021] The construction of the target influence matrix is based on the multi-objective correlation analysis framework. The row vectors of the matrix represent the optimization objectives, and the column vectors reflect the types of interaction between the objectives. There is a synergistic relationship between energy consumption optimization and transportation path optimization, and shortening the transportation distance directly affects fuel consumption. There is a restrictive relationship between equipment efficiency optimization and material turnover rate optimization. Centralized processing of steel reinforcement speeds up equipment efficiency but increases site storage pressure. The synergistic strength value is obtained by statistical analysis of historical data, and the correlation curve of the simultaneous target achievement is analyzed. The conflict frequency monitoring method is used to calculate the restrictive coefficient, and the probability of target conflict events in the construction log is recorded. The matrix real-time updating mechanism automatically captures the relationship changes in the new construction phase. When the main structure construction period is converted to the decoration stage, the priority of the material transportation target is automatically adjusted.

[0022] The parameter sensitivity analysis uses a dynamic threshold triggering mechanism. The change amplitude of the adjustable parameters is continuously monitored, and when the fluctuation value exceeds the preset threshold, the sensitivity coefficient calculation is started. The calculation method integrates the direct influence degree of the main target and the indirect influence degree between the auxiliary targets. The former is determined by the function gradient of the parameter change on the main optimization target, and the latter introduces a transmission factor to calculate the secondary effect of the parameter on the main target through the auxiliary target. The sensitivity calculation of the tower crane hoisting speed parameter includes three dimensions: the contribution degree of the main target to directly shorten the construction period, the transmission influence through the auxiliary target to reduce fuel consumption, and the restrictive correction factor during equipment collaborative operation. The calculation results form a parameter sensitivity spectrum diagram, with the horizontal axis as the parameter identifier and the vertical axis as the multi-objective comprehensive sensitivity coefficient.

[0023] The key monitoring parameter screening executes a hierarchical screening strategy. The first level selects the top six parameters of the main target sensitivity coefficient as the core monitoring items, and the second level includes the parameters with auxiliary target transmission coefficients exceeding the critical value as supplementary monitoring items. The dynamic screening module is associated with the construction state recognition system. During the foundation pit support stage, the dewatering equipment operation parameters are automatically included, and during the upper structure construction, the concrete vibrator energy consumption parameters are switched. The final output of the key monitoring parameter list is attached with a double-dimensional index: the main target influence strength value is presented in a column chart, and the multi-objective coupling influence value is displayed in a chromatogram. The analysis report generation module automatically compares the deviation degree of the current parameter set and the historical optimal set, and marks the sensitive parameter abnormal fluctuation state. The parameter tuning suggestion unit combines the sensitivity ranking to output the parameter adjustment priority list and the relationship mapping table of the expected target improvement direction.

[0024] The monitoring strategy update instruction is automatically generated based on the analysis result. When the sensitive coefficient of the material transportation path parameter jumps, the vehicle-mounted GPS positioning sampling frequency adjustment instruction is triggered, and the transportation fuel consumption monitoring points are increased synchronously. The sensitive parameter fluctuation of the equipment operation efficiency triggers the video analysis system linkage, and the idle time of the equipment is automatically calculated through image recognition. The strategy update package is pushed in the form of a digital work order, including parameter monitoring threshold value update table, data acquisition frequency adjustment scheme and target weight redistribution matrix. The entire analysis process is completed in real time on the edge computing node, and the analysis cycle is synchronized with the construction progress plan, and a dynamic optimization target influence analysis report is automatically output at each daily construction node.

[0025] Embodiment 2: refer to Figure 3 In the process of dynamic monitoring of carbon emissions in building construction, parameter dynamic optimization is realized. The process starts from establishing the control threshold interval of key monitoring parameters. The key monitoring parameters are determined by the influence analysis of the previous stage, covering categories such as equipment operation, material management and energy distribution. The setting of the equipment power fluctuation range is based on the technical specifications of the equipment model, such as the rated power curve and safe operation specifications of the concrete pumping equipment, and the confidence interval is obtained by statistical analysis of historical operation data. The boundary conditions of the material transportation time window are determined in combination with the construction progress plan, considering traffic rules, weather factors and material properties, and defining the allowed minimum and maximum transportation period. The boundary conditions are stored in the form of interval values in the constraint database and are associated with real-time monitoring equipment to capture the parameter fluctuation range of the current construction state. The control threshold is loaded at system initialization and automatically updated as the construction stage changes, for example, increasing the power upper limit of the vibrator in the pouring stage. The dynamic interval adjustment module periodically scans the parameter deviation value, and triggers the threshold recalibration program when consecutive multiple critical values are detected.

[0026] A multi-objective particle swarm optimization algorithm is used to create an initial solution set, each solution representing a complete parameter combination scheme. The structure of the solution is based on key monitoring parameters, such as a combination of variables including tower crane lifting speed, formwork turnover period and fuel distribution ratio. The particle swarm initialization process imports the control threshold interval as the boundary constraint of the solution space, uniformly generates particle positions and assigns initial velocities. The number of particles is adaptively set according to the parameter dimension, with the rule being twice the number of dimensions to ensure diversity coverage. The initial velocity vector is generated in a random direction, and the amplitude is related to the historical average change rate of the parameter. The solution set is stored in distributed memory, indexed by particle identifier for subsequent iteration access. The spatial distribution matrix of the solution is automatically constructed to record the relative position relationship between particles. The creation of the initial solution set is bound to the real-time construction data stream, and corresponding parameter particles are automatically added when a new construction area starts. The solution set update mechanism checks invalid position particles, and particles that exceed the constraint range trigger position reset logic to maintain the solution set within the feasible solution space. The evaluation period of the initial solution set is synchronized with the construction node, and a new solution generation sequence is automatically executed at the end of each daily plan.

[0027] The fitness evaluation function is based on a balanced adaptation model, integrating carbon emission intensity, construction efficiency, and economic cost as three-dimensional indicators. The carbon emission intensity index is calculated using the equivalent conversion method, which maps the energy consumption, material loss, and equipment emission data captured by monitoring equipment into standard carbon emission values. The construction efficiency index integrates progress plan completion rate and quality acceptance data, and the comprehensive efficiency value is obtained by weighted summation. The economic cost item covers equipment depreciation, labor consumption, and material waste cost, based on real-time data input from the financial system. The evaluation function calculates the fitness score through multi-dimensional weighting, and the weight coefficients are derived from the characteristics of the construction phase. The evaluation process traverses each particle in the solution set, calling the real-time database to obtain the index data corresponding to the current parameter combination. The fitness calculator executes in parallel, with each particle allocated an independent computing thread to shorten the evaluation time. The calculation results form a particle fitness distribution graph, with the horizontal axis representing particle number and the vertical axis representing multi-dimensional score value. The particle state monitor runs synchronously, detecting the distribution characteristics of low fitness particles. The fitness evaluation is linked with the threshold interval, and the score is automatically reduced when the particle parameters exceed the boundary, maintaining the compliance of constraints. The periodic re-evaluation program handles sudden changes in working conditions, such as triggering a re-score when equipment fails.

[0028] The parameter optimization constraint conditions integrate process limitations, safety specifications, and environmental requirements to build a constraint knowledge base. Process constraints include hard limits on concrete initial setting time for vibration frequency, safety constraints involve mandatory spacing for equipment safety distance, and environmental constraints handle noise control indicators. The constraint library extracts rules from construction specification documents, defining infeasible regions with logical expressions. The optimization direction is determined based on constraint conditions and fitness results, and the direction calculation uses a group guidance strategy to identify the aggregation direction of the high fitness particle group as the main search direction. The step size calculation module analyzes the fitness difference gradient between adjacent particles, adjusting the search step size through the difference gradient. Large differences trigger large step jumps to cross local optima. The direction-step size mapping table is generated in real time, including direction identifiers, step size values, and constraint compatibility check markers. The direction update program is executed before each iteration, scanning constraint library changes to adjust direction vectors. The direction guides the particle velocity update rule, such as increasing the attraction strength between particles when the group moves towards the high-efficiency parameter area. The step size adaptive mechanism monitors particle movement effects, automatically enlarging the step size when there are too many stagnant particles.

[0029] The iterative process adopts multiple rounds of evolutionary cycles, and each round of iteration records the optimized trajectory of each particle solution. The trajectory data structure includes a historical position sequence, a fitness change sequence, and a timestamp sequence. The recording point density is configurable, and a recording point is captured every certain number of steps. The trajectory storage and index use a high-concurrency database to support the simultaneous recording of multiple particles. The trajectory analyzer is executed at the iteration interval, extracting the path features of each particle. The evolution state of the solution is monitored through a sliding window mechanism, and the window size defines a dynamic window strategy based on the number of iterations, with the initial window covering the recording points of the first three iterations. The state recognition engine runs at the end of each window period, analyzing the fitness change trend, position movement amplitude, and group distance of the particles. The state classifier automatically determines the particle state type according to the recognition results. The trend feature extraction module calculates the derivative sign and amplitude of the fitness within the window, and the position analysis module evaluates the drift pattern of the particle in the solution space. The state output includes a continuous stable evolution identifier.

[0030] The trajectory classification logic divides the iterative trajectory into three categories: optimal solution trajectory, exploration trajectory, and inferior solution trajectory. The optimal solution trajectory is determined by continuous multiple window fitness growth and smooth position movement, indicating stable proximity to the optimal region. The identification of the exploration trajectory is based on frequent directional changes and moderate fitness oscillation characteristics, indicating new region detection. The definition of the inferior solution trajectory is based on continuous fitness decline or stagnation state, reflecting ineffective search. The classification result is marked by a trajectory type identifier to indicate the particle state. The strategy allocation module performs differential processing according to the trajectory type. The optimal solution trajectory adopts an elite preservation strategy, creating an elite subset by copying high fitness particles, and the elite number ratio is dynamically set based on group density. Elite particles participate in subsequent iterations with conservative update rules, limiting the movement amplitude. The exploration trajectory applies a diversity preservation mechanism, adjusting the distance between particles through repulsive forces to prevent clustering, such as introducing directional perturbations in particle dense areas. The inferior solution trajectory triggers the re-initialization logic, resetting the position to a random feasible point and restoring the initial velocity when the particle is in the inferior solution state for multiple windows. Strategy execution threads are processed with priority, and the optimal solution trajectory strategy is applied immediately after each round of iteration.

[0031] The whole optimization cycle adopts a step-by-step execution mode. After the initial solution set evaluation, the iteration main loop is entered. The particle position update logic is executed at the beginning of each iteration, and the particle movement is adjusted according to the optimization direction step size. After the particle movement, the new fitness is calculated and the trajectory record is updated. The sliding window monitors the analysis state at the end of the iteration and classifies the trajectory. The strategy application step distributes different strategy tasks according to the classification results. The loop exit condition is based on the iteration number threshold or the fitness stability test. The final output parameters of the dynamic optimization scheme include the optimized parameter combination set and the matching monitoring strategy update package. The scheme generator summarizes the parameter values of the elite particles as the recommended scheme and outputs it in a standard data format. The scheme deployer is integrated into the monitoring system to automatically push the update instructions. The data flow is synchronized in real time with the construction state during the processing. The monitoring log records all operation events to support audit tracing. The system fault tolerance processing includes iteration interruption recovery and particle loss reloading functions, maintaining the continuous and stable optimization.

[0032] Example 3: see Figure 4 In the optimal solution trajectory classification, the local prediction model is established by using the time series analysis method to process the historical optimization data. The model input is the parameter sequence of the particle wherein represents the parameter vector at the th iteration, is the time window size. The prediction output is the improvement trend vector , and the component represents the expected change direction of each parameter. The trend predictor adopts a sliding window training mechanism, and when the number of data points in the window is not less than 50, the model is updated. The prediction result generates a reference optimization direction , and the direction angle calculation is based on the principal component analysis of the historical improvement path in the parameter space. The direction correction module monitors the prediction deviation in real time, and when the prediction error exceeds the threshold for three consecutive times, the model is retrained. The particles in the optimal solution trajectory adopt a fine-tuning iteration update strategy, and the update step size is positively correlated with the prediction confidence, which is calculated by the inverse ratio of the prediction residual. The fine-tuning process applies parameter coupling constraints, for example, when adjusting the concrete pouring speed, the change range of the vibration frequency is also limited.

[0033] The penalty recognition layer configured in the poor solution trajectory classification adopts an error pattern detection algorithm. The layer input is a particle state triple , wherein contains parameter values and fitness values. The error direction identifier calculates the angle between the state change vector and . When and , it is marked as an error improvement direction. The window improvement taboo table stores the recently detected error direction patterns, and the table entry format is wherein is the error parameter combination feature, is the error moving direction, is the entry validity period. The taboo checker scans the similarity between the current moving direction and the taboo entries before particle update, and triggers direction correction when the similarity exceeds the threshold. The taboo learner analyzes the distribution regularity of the newly added taboo entries and automatically adjusts the detection sensitivity parameters. The cleaning period of the taboo table is synchronized with the construction phase transition, and all entries are reset at the beginning of the new phase.

[0034] The exploration trajectory classification adopts a hybrid exploration strategy, combining the reference optimization direction and the taboo constraint. The direction synthesizer calculates the final exploration direction wherein is the direction mixing coefficient, which is dynamically adjusted based on the historical exploration effect of the trajectory. The random direction component is generated by considering the parameter sensitivity distribution, and high sensitivity parameters obtain a higher probability of small perturbation. The exploration step size adopts an adaptive mechanism, with the initial step size set to 5% of the parameter range, and subsequently adjusted according to the exploration effect index. The exploration effect index calculates the average improvement rate of fitness of the last three explorations, and when the step size is reduced to times the original value. The update process of the exploration trajectory applies a diversity preservation constraint, preventing the over-concentration of parameter combinations through inter-particle repulsive force. The repulsive force calculation is based on the similarity measure in the parameter space, and particles with a similarity exceeding the threshold trigger mutual repulsion velocity components.

[0035] The random perturbation update of the inferior solution trajectory adopts a dimension-by-dimension processing strategy. Gaussian perturbation is applied to high sensitivity parameters wherein takes 2% of the parameter allowed range. Uniform perturbation is adopted for moderately sensitive parameters , with a step size of 5% of the parameter range. Low sensitivity parameters implement directional perturbation, with the direction determined by the relative position of the nearest neighbor optimal solution particle. The perturbation amplitude is negatively related to the iteration number, allowing larger amplitude jumps in the early stage and gradually converging in the later stage. The perturbed parameter combination is verified for feasibility by the constraint checker, and perturbations that violate constraints trigger a retry mechanism with a maximum of 10 retries. The perturbation recorder tracks the perturbation history of each parameter, and the effective perturbation direction is used to optimize the perturbation strategy.

[0036] The construction carbon intensity evaluation function is constructed using a hierarchical calculation framework. The basic layer handles direct emission sources, including equipment fuel consumption , construction electricity consumption , and material processing loss . The intermediate layer calculates transportation emissions and temporary facility emissions Top-level integration implies carbon emissions. This includes carbon emissions from building materials. The function expression is:

[0037] in: to The weighting coefficients for each emission source are determined using life cycle assessment methods. The variables are the standard emissions after unit conversion of the raw data collected by the monitoring equipment. The function calculator receives data stream input in real time and updates the calculation results every 15 minutes. The weight adjustment module responds to changes in the construction phase, such as increasing the weight during the structural construction phase. Weighting, improvement during the renovation stage importance.

[0038] Construction efficiency evaluation function Integrating the time dimension and quality dimensions Time evaluation is based on schedule deviation rate calculation, taking into account the completion status of critical path tasks. Quality evaluation collects acceptance pass rate data, combined with the cost of quality defect repair. The function uses normalization processing, and the output value range is [0,1]. The efficiency calculation engine interfaces with the project management system to automatically obtain schedule updates and quality reports. The abnormal data processing module identifies and corrects input anomalies; for example, sudden schedule pause events are marked as special states and temporarily excluded from calculation.

[0039] Balanced fitness function The construction uses a dynamic weighting method. The basic form is:

[0040] in: The initial value for the balancing coefficient is set to 0.6. The coefficient adjuster monitors the function output distribution of the most recent evaluation period. or Automatic adjustment when the proportion of extreme values ​​is too high Value. The function optimizer periodically searches for the optimal combination of coefficients, testing different values ​​using a grid scan method. The overall evaluation effect is determined by the value. The function output is standardized to a percentage score for easy comparison across different schemes.

[0041] The optimization scheme screening implements a multi-stage filtering mechanism. The preliminary screening phase excludes schemes that violate hard constraints, such as parameter combinations not allowed by safety regulations. The fine screening phase is based on ranking by balanced fitness scores, retaining the top 20% of candidate schemes. The final selection phase introduces a human decision-making interface, combining the technical parameters of the preferred scheme with a visual construction simulation for display. The scheme report generator automatically creates a complete document containing the basis for parameter settings, expected effects, and risk warnings. The scheme deployer pushes the final selected parameter combination to the monitoring system, while updating control thresholds and alarm rules. The version control system records all candidate schemes and selection criteria, supporting scheme backtracking and effect comparison analysis.

[0042] Embodiment 4: Referring to Figure 5 The project includes a reinforced concrete frame-core tube structure with 2 underground floors and 25 above-ground floors. The spatiotemporal feature segmentation module collects carbon emission intensity data for four construction areas during the main structure construction phase, spanning from March to August 2023. The data collection frequency is set to record every hour, achieved through an Internet of Things sensor network installed at key locations such as tower cranes, concrete pump trucks, and steel processing areas. After cleaning, the raw data is stored in a time series database, with each data point containing four fields: timestamp, area number, device type, and carbon emission. A segment of carbon emission intensity data for the underground first floor area on a typical construction day is shown in Table 1.

[0043] Table 1: Segment of carbon emission intensity data for the underground first floor area on a typical construction day.

[0044]

[0045] When processing the above time series data, the feature mutation detection algorithm first calculates the first-order difference values of adjacent time points. The algorithm sets the sliding window size to 24 hours, and marks a potential mutation point when there are three consecutive difference values exceeding the threshold of 8 kgCO2 / h within the window. In the construction log of April, the system detects a mutation event at 10:45 on April 15, corresponding to the process conversion from steel processing to concrete pouring on site. After manual review and confirmation, this time is recorded as the official segmentation point, dividing the morning steel operation phase from the afternoon concrete operation phase. A total of 37 valid feature mutation points are identified for the entire project, with an average of one significant process conversion occurring every 4.3 days.

[0046] The criterion for determining the stable stage of construction is that the fluctuation range of carbon emissions within a certain duration is less than 15%, and the main construction activity type remains unchanged. In this project, the standard floor construction presents typical stable characteristics, such as the carbon emission intensity during the 7th floor slab construction period (May 6-9) maintaining at 22±3 kgCO2 / h, and the main equipment combination being 2 tower cranes and 1 concrete pump truck operating continuously. The construction of the monitoring chain tree structure takes the equipment node as the core, for example, the tower crane node is subdivided into three sub-nodes: hoisting mechanism, slewing mechanism, and luffing mechanism, each of which is associated with corresponding energy consumption monitoring data.

[0047] The hierarchical optimization strategy is implemented in a three-level architecture in this project. The bottom layer optimization targets the parameter setting of a single tower crane, including the optimization of the ratio of lifting speed to standby time, which reduces energy consumption by 5% by adjusting the motor frequency. The middle layer optimization coordinates the collaborative work of the tower crane and the concrete pump truck, and dynamically adjusts the pumping rhythm based on real-time tracking of the hopper position by RFID technology, reducing the equipment waiting time by 18%. The high-level optimization coordinates the entire standard floor construction flow section division, adjusting the original design of 4 flow sections to 5, making the tower crane utilization rate curve tend to be stable. The optimization instructions are issued through the construction management platform, and the execution results are fed back to the monitoring system to form a closed loop.

[0048] The typical scenario of the construction change stage appears during the construction of the transfer floor (16th floor), which changes the structure form, causing the failure of the conventional equipment combination. The dynamic monitoring data flow system captures the abnormal emission peak at 14:00 on May 28, with the concrete pump truck emitting 35 kgCO2 per hour, which is 40% higher than the normal range. The system automatically triggers the data flow analysis channel to compare the current equipment operating parameters with the historical optimal mode in real time. The analysis finds that the pump truck pressure setting is too high, which is due to the increase in the thickness of the transfer floor wall without timely adjustment of the equipment parameters. The benchmark value provided by the static monitoring model is 28 kgCO2 / h, and the dynamic calibration module gives a reasonable threshold of 32 kgCO2 / h after considering the actual working conditions, and pushes the pressure parameter adjustment suggestion to the equipment control system.

[0049] The node weight update of the monitoring chain tree structure occurs during the steel structure hoisting stage. The original node weight is set based on the concrete structure construction data, with the tower crane node weight accounting for 70%. When steel column hoisting begins in June, the system detects that the node contribution degree distribution is abnormal, with the actual carbon emission proportion of the mobile crane node rising to 45%. The weight redistribution algorithm recalculates the influence coefficient of each node based on the monitoring data of the previous 72 hours in the new stage. The updated tree structure increases the weight of the mobile crane node to 40% and adds a sub-node for the high-strength bolt tightening equipment.

[0050] The spatiotemporal feature analysis is also applied to the construction section division optimization in this project. By analyzing the spatiotemporal distribution characteristics of carbon emissions in each region, it is found that the original design of parallel construction in the east and west regions has equipment resource conflicts. The system suggests changing to a staggered construction mode, with the east region focusing on vertical transportation operations and the west region arranging ground material preparation. After adjustment, the equipment interference between regions is reduced, and the tower crane collision warning events are reduced by 63%. The feature analysis report automatic generation module regularly outputs the spatiotemporal thermal map of carbon emissions in each region, assisting managers to intuitively grasp the construction intensity distribution.

[0051] The prediction function of construction phase transition is based on historical mutation point data. The system analysis found that the emission decreased usually 2 hours before the concrete pouring ended, accordingly, the system gave an early warning that the floor construction phase would end on May 20th. The prediction information triggered the material preparation instruction, so that the subsequent masonry work started 8 hours earlier. The prediction model continuously learns new data, and its accuracy gradually improves as the project progresses, with the prediction error of phase transition controlled within 2 hours in the later stage. During the entire implementation process, all monitoring data and analysis results are integrated into the digital twin platform, realizing dynamic three-dimensional visualization monitoring of carbon emissions.

[0052] Example 5: Illustrated by a case of a highway bridge reconstruction project. The project involves the intersection of old bridge demolition and new bridge construction, and the construction site contains six dynamically changing regions. The dynamic monitoring data flow injection mechanism is realized through distributed data collection terminals. The terminal equipment includes three types of sensor networks: vibration and oil consumption sensor units attached to engineering machinery, weight-humidity composite sensors deployed in material storage yards, and GPS trajectory recorders deployed on transportation paths. The data flow transmission uses a hybrid network of 5G and LoRa, with a sampling frequency of 2 times per second for key parameters and event-triggered collection for non-key parameters. The data flow processing channel includes a three-level cache structure, with raw data flow first entering the edge computing node for preliminary filtering, then transmitted to the regional processing server for feature extraction, and finally aggregated to the central monitoring platform.

[0053] In the old pier column demolition phase, the static monitoring model establishes a benchmark value based on historical blasting demolition data. The model input features include concrete strength grade , steel density and structure volume , and the output is the predicted blasting equivalent coefficient and the predicted dust emission. When the No. 3 pier was demolished on September 18, 2023, the real-time data flow showed that the vibration frequency of the drilling equipment was consistently lower than the model's expected value by 12±2 Hz. The collaborative analysis engine calculates the deviation index between the dynamic measured value and the static predicted value:

[0054] where: To monitor the number of parameters, denotes the dynamic acquisition value, is the static model prediction value, is the historical data standard deviation. When the system automatically triggers the real-time calibration program for 10 minutes.

[0055] The parameter calibration process adopts an improved Kalman filter algorithm. The state vector contains key monitoring parameters: drilling depth , blast hole spacing , charge weight . The observation vector comes from dynamic sensor data. The process noise matrix of the filter is dynamically adjusted according to the state of the equipment, such as increasing the corresponding element value of when the drill bit wear coefficient .

[0056] The node weight update mechanism of the monitoring chain tree structure is designed based on the time factor. The tree structure contains three types of nodes: equipment nodes (crushing hammer, crane, etc.), process nodes (cutting, lifting, etc.), and area nodes (eastern span of old bridge, western span, etc.). The initial weight is set by engineering characteristics. The weight update function introduces a time decay factor , where denotes the daily weight decay rate, is the initial weight setting time. During the demolition of the western span box girder, the initial weight of the automobile crane node is 0.25, which decreases to 0.22 after three days of construction according to the decay formula. After inputting the dynamic calibration results, the node contribution is recalculated:

[0057] In the formula: denotes the change amplitude of the node-related monitoring parameters, is the parameter influence coefficient, represents the node operation duration factor. The weight recalculation on September 20th increases the crushing hammer node weight from 0.31 to 0.35, reflecting its key role in concrete breaking. The node topology relationship is updated synchronously, and a new "hydraulic shear" node is added as a sub-node of the "reinforcing steel cutting" process, with an initial weight set to 0.08.

[0058] The aging weight control module implements a four-level management strategy. The first-level node performs automatic weight update every 8 hours, the second-level node updates according to the construction phase, the third-level node updates daily, and the fourth-level node adopts a fixed weight. The update event triggering strategy includes timing triggering (shift time point), event triggering (process conversion), and abnormal triggering (monitoring deviation alarm). The weight update log records the basis data for each adjustment in detail, forming a complete weight evolution map for traceability analysis.

[0059] The dynamic compensation subsystem plays a significant role during rainstorm weather. When the meteorological monitoring module forecasts rainfall intensity exceeding 20 mm / h, the system activates the wet work compensation program. The weight of the concrete pouring node is automatically increased by a coefficient , wherein is the real-time rainfall intensity. At the same time, the vibration rod energy consumption parameter threshold is dynamically adjusted, allowing a 10% deviation floating range. The monitoring strategy is synchronously switched to a high-frequency mode, and the key parameter sampling rate is increased to 5 times per second.

[0060] The continuous evolution of the dynamic model is realized through an incremental learning mechanism. After completing each construction change phase, the system automatically extracts a feature data package , including the number of calibration events , the effective adjustment proportion , and the parameter convergence curve . The feature package is input into the model evolutioner, which adjusts the model parameter sensitivity coefficient by comparing the difference between the actual effect and the expected effect . The difference is calculated using the sliding window variance analysis method, and the model structure optimization is triggered when three consecutive windows . All dynamic processing processes are recorded through the blockchain notarization system, and each data change event generates an independent time stamp and hash value, ensuring the integrity and auditability of the monitoring data.

[0061] It should be noted that, in the present document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0062] While embodiments of the present application have been shown and described with reference to particular embodiments thereof, it will be understood by those skilled in the art that various changes in form and details can be made therein without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. A method for dynamic monitoring of carbon emissions in construction, characterized by, The method comprises: acquiring carbon emission activity data of the building construction process, and establishing a carbon emission dynamic monitoring database mapped with the current construction stage; configuring a dynamic optimization target according to the carbon emission dynamic monitoring database, the dynamic optimization target comprising energy consumption optimization, material turnover rate optimization, equipment operation efficiency optimization, and transportation path optimization; determining a target classification identifier in the dynamic optimization target, the target classification identifier comprising a main optimization target, an auxiliary optimization target, and a benchmark maintenance target; after selecting key monitoring parameters using the main optimization target and the auxiliary optimization target, performing dynamic optimization target influence analysis on the key monitoring parameters; after establishing parameter optimization constraints according to the dynamic optimization target influence analysis result, performing parameter dynamic optimization, and updating the carbon emission dynamic monitoring strategy using the parameter dynamic optimization result.

2. The method of claim 1, wherein, The dynamic optimization target influence analysis on the key monitoring parameters after selecting the key monitoring parameters using the main optimization target and the auxiliary optimization target comprises: acquiring a set of adjustable parameters of the building construction process; quantitatively mapping the influence degree of the set of adjustable parameters on the main optimization target and the auxiliary optimization target; constructing a target influence matrix of all optimization targets in the dynamic optimization target, the target influence matrix representing the mutual relationship between different optimization targets, the mutual relationship comprising a positive synergistic relationship and a negative constraint relationship; calculating the sensitivity coefficient of the set of adjustable parameters according to the influence degree quantitative mapping and the target influence matrix; establishing a dynamic optimization target influence analysis result according to the sensitivity coefficient calculation result.

3. The method of claim 2, wherein the construction carbon emission dynamic monitoring method is characterized by, The parameter dynamic optimization comprises: after establishing a control threshold interval of the key monitoring parameters, creating an initial solution set based on the current parameter set; after performing solution fitness evaluation in the initial solution set, establishing an optimization direction and an optimization step length through parameter optimization constraints and fitness evaluation results; performing initial solution set iterative updating using the optimization direction and the optimization step length; outputting a parameter dynamic optimization scheme according to the iterative updating result.

4. The method of claim 3, wherein the construction carbon emission dynamic monitoring method is characterized by, The initial solution set iterative updating using the optimization direction and the optimization step length comprises: establishing an iteration trajectory for each solution, and identifying the iteration trajectory through the solution fitness value of each iteration; configuring an iteration evaluation window, identifying the update state of the iteration trajectory in the iteration evaluation window, generating a trajectory classification identifier, and the trajectory classification identifier comprising a good solution trajectory classification, an exploration trajectory classification, and a poor solution trajectory classification; performing search self-optimization management of the iteration updating according to the trajectory classification identifier.

5. The method of claim 4, wherein the construction carbon emission dynamic monitoring method is characterized by, The search self-optimization management of the iteration updating according to the trajectory classification identifier comprises: configuring a local prediction model in the good solution trajectory classification, performing improved trend prediction using the local prediction model, and generating a first reference optimization direction; configuring a penalty identification layer in the poor solution trajectory classification, identifying an error improvement direction using the penalty identification layer, and establishing a window improvement taboo; Performing iterative update of the solution within the solution trajectory classification with the first reference optimization direction, window improvement tabu, performing mixed exploration iterative update of the exploration trajectory classification with the first reference optimization direction, window improvement tabu, and configuring a random disturbance factor to perform iterative update of the solution within the inferior solution trajectory classification.

6. The method of claim 1, wherein, The dynamic parameter optimization further comprises: Establishing a carbon emission intensity evaluation function of the monitoring parameter; Establishing a balance adaptation function according to the carbon emission intensity evaluation function and the construction efficiency evaluation function; Performing optimization scheme screening of the parameter dynamic optimization based on the balance adaptation function, and outputting the optimization scheme screening result as the parameter dynamic optimization result.

7. The method of claim 1, wherein, The method further comprises: Dividing the construction stable stage and the construction change stage based on a space-time feature segmentation method; Constructing a monitoring chain tree structure of the device node and the process edge within the construction stable stage; Establishing a hierarchical optimization strategy according to the monitoring chain tree structure.

8. The method of claim 7, wherein the construction carbon emission dynamic monitoring method is characterized by, The method further comprises: Obtaining time series carbon emission intensity data of different construction areas; Calculating a first-order difference sequence of the time series carbon emission intensity data, and screening a feature mutation time according to the first-order difference sequence; Dividing the construction stable stage and the construction change stage by taking the feature mutation time as a segmentation point.

9. The method of claim 8, wherein, The method further comprises: Injecting a dynamic monitoring data stream in the construction change stage; Performing real-time calibration of the key monitoring parameter based on a cooperative analysis mechanism of the static monitoring model and the dynamic monitoring data stream; Updating the node weight of the monitoring chain tree structure according to the real-time calibration result.

10. A dynamic monitoring system for carbon emissions in construction, characterized by, The system for implementing the building construction carbon emission dynamic monitoring method of any one of claims 1-9, the system comprises: A data acquisition module for acquiring carbon emission activity data of the building construction process and establishing a carbon emission dynamic monitoring database mapped with the current construction stage; A target configuration module for configuring a dynamic optimization target according to the carbon emission dynamic monitoring database, the dynamic optimization target comprising energy consumption optimization, material turnover rate optimization, equipment operation efficiency optimization, and transportation path optimization; A classification identification module for determining a target classification identification in the dynamic optimization target, the target classification identification comprising a main optimization target, an auxiliary optimization target, and a benchmark maintenance target; An influence analysis module for performing dynamic optimization target influence analysis of the key monitoring parameter after selecting the key monitoring parameter by using the main optimization target and the auxiliary optimization target; A dynamic optimization module for performing parameter dynamic optimization after establishing parameter optimization constraints according to the dynamic optimization target influence analysis result, and updating the carbon emission dynamic monitoring strategy by using the parameter dynamic optimization result.

Images