A full life cycle environmental protection engineering project management system and method

Abstract

This invention discloses a full life-cycle environmental protection engineering project management system and method, relating to the field of engineering project management technology. It includes the following steps: acquiring design parameters from the project initiation phase, measured parameters from the construction phase, and monitoring parameters from the operation and maintenance phase; synchronously mapping the parameters of each phase according to a time axis to generate a multi-dimensional project data map containing timestamps; based on the multi-dimensional project data map, identifying a first difference feature between the design parameters from the initiation phase and the measured parameters from the construction phase, and identifying a second difference feature between the measured parameters from the construction phase and the monitoring parameters from the operation and maintenance phase; inputting the first and second difference features into a pre-constructed deviation transmission analysis model to calculate the cumulative impact weight of the design deviation in the construction phase and the attenuation coefficient of the construction deviation in the operation and maintenance phase.

Description

Technical Field

[0001] This invention relates to the field of engineering project management technology, specifically to a full life-cycle environmental protection engineering project management system and method. Background Technology

[0002] Engineering projects generate a large amount of heterogeneous data at each stage, from project initiation and construction to operation and maintenance. Existing technical solutions typically manage data independently for each stage, lacking mechanisms for cross-stage data fusion and correlation analysis. This fragmented data management approach makes it impossible to effectively track the actual transmission of design intent during the construction and operation and maintenance phases. When baseline parameters established in the design phase deviate in subsequent stages, project managers find it difficult to accurately identify the source and evolution path of the deviation.

[0003] Existing technical solutions, when discovering discrepancies between measured construction values ​​and design values, or deviations in operation and maintenance monitoring values ​​from expectations, often only allow for localized adjustments at the current stage. How deviations in the construction phase gradually accumulate and propagate to the operation and maintenance phase through technological influences, and how deviations in the operation and maintenance phase are attenuated by environmental factors, cannot be quantitatively expressed. The lack of quantitative modeling of the deviation propagation process makes it impossible to determine whether significant deviations in monitoring parameters during the operation and maintenance phase are caused by the accumulation of early design deviations or by independent environmental factors in later stages, thus hindering the decision on whether to make corrective adjustments to the design itself. This lack of deviation tracing capability becomes a major obstacle to closed-loop management throughout the entire lifecycle of engineering projects.

[0004] One problem this solution needs to address is how to structurally represent the heterogeneous parameters of the three phases of project initiation, construction, and operation and maintenance on a unified time dimension, so that the parameter differences between phases can be accurately located and encoded. Another problem is how to perform hierarchical modeling of the cumulative effect of design deviations after being affected by construction technology and the attenuation effect of construction deviations after being affected by environmental factors, and to feed the cumulative impact back to the attenuation calculation process to obtain a quantitative transmission relationship that can support reverse correction. Summary of the Invention

[0005] The purpose of this invention is to provide a full life-cycle environmental protection engineering project management system and method. By performing cross-stage correlation and deviation transmission modeling on the full life-cycle data of the engineering project, it realizes the quantitative analysis of the cumulative effect of design deviation and the attenuation effect of construction deviation, and then, based on the analysis results, reversely corrects the operation and maintenance benchmark and drives the design correction decision.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention discloses a method for managing environmental protection engineering projects throughout their entire life cycle, comprising: acquiring design parameters during the project initiation phase, measured parameters during the construction phase, and monitoring parameters during the operation and maintenance phase; synchronously mapping the parameters of each phase according to a time axis to generate a multi-dimensional project data map containing timestamps; based on the multi-dimensional project data map, identifying a first difference feature between the design parameters during the project initiation phase and the measured parameters during the construction phase, and identifying a second difference feature between the measured parameters during the construction phase and the monitoring parameters during the operation and maintenance phase; inputting the first and second difference features into a pre-constructed deviation transmission analysis model to calculate the cumulative impact weight of the design deviation during the construction phase and the attenuation coefficient of the construction deviation during the operation and maintenance phase; based on the cumulative impact weight and the attenuation coefficient, performing reverse correction on the monitoring parameters during the operation and maintenance phase to generate corrected operation and maintenance benchmark parameters, and comparing the corrected operation and maintenance benchmark parameters with the design parameters during the project initiation phase to determine whether a project design correction instruction needs to be triggered. Through the above solution, this invention can connect the data chain of the entire life cycle of an engineering project from design and construction to operation and maintenance, accurately depict the transmission and evolution of deviations at different stages, and then deduce the reasonable construction benchmark corresponding to the current operation and maintenance status, providing a quantitative basis for the continuous optimization of the design scheme.

[0007] As a preferred technical solution of the present invention, the process of obtaining design parameters in the project initiation stage, measured parameters in the construction stage, and monitoring parameters in the operation and maintenance stage of an engineering project, and synchronously mapping the parameters of each stage according to the time axis to generate a multi-dimensional project data map containing timestamps includes: extracting structural design parameters and material design parameters from the design documents in the project initiation stage, extracting measured structural parameters and measured material parameters from the on-site records in the construction stage, and extracting structural response parameters and environmental action parameters from the monitoring system in the operation and maintenance stage; assigning a corresponding timestamp to each extracted parameter, which records the design completion time in the project initiation stage, the measured record time in the construction stage, and the monitoring acquisition time in the operation and maintenance stage; establishing a two-dimensional mapping table with time as the horizontal axis and parameter type as the vertical axis, filling the design parameters, measured parameters, and monitoring parameters with timestamps into the two-dimensional mapping table, and associating and labeling the values ​​of the same parameter type under different timestamps to form a multi-dimensional project data map. In this way, the engineering parameters that were originally scattered in different stages and carriers are unified and integrated into a structured map with time and parameter type as the dimensions. The values ​​of the same parameter in different life stages are unfolded and correlated along the time axis, which facilitates cross-stage comparison and deviation identification directly in the same framework.

[0008] As a preferred technical solution of the present invention, the process of identifying the first difference feature between design parameters in the project initiation stage and measured parameters in the construction stage, and identifying the second difference feature between measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage based on a multi-dimensional project data map, includes: extracting the values ​​of design parameters in the project initiation stage and measured parameters in the construction stage one by one according to parameter type in the multi-dimensional project data map; calculating the difference and ratio between the same type of design parameters and measured parameters; and encoding the difference and ratio as the first difference feature; extracting the values ​​of measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage one by one according to parameter type; calculating the difference and ratio between the same type of measured parameters and monitoring parameters; and encoding the difference and ratio as the second difference feature; and normalizing the first difference feature and the second difference feature respectively, so that the dimensions of the first difference feature and the second difference feature are unified to a unitless scale. By incorporating both the absolute difference and the relative ratio into the feature encoding and normalizing them, the absolute magnitude of the deviation is preserved, while also reflecting the relative degree of the deviation. This eliminates the incomparability caused by differences in units and orders of magnitude between different parameter types, enabling subsequent models to handle deviations of various parameter types on a uniform scale.

[0009] Preferably, when normalizing the first and second difference features, the range normalization method is used to map the difference and ratio to the [0,1] interval and then re-encode them. This can effectively suppress the influence of extreme values ​​on the feature distribution and ensure the stable expression of feature values ​​within a limited range.

[0010] As a preferred embodiment of the present invention, the process of inputting the first difference feature and the second difference feature into a pre-constructed deviation transmission analysis model to calculate the cumulative impact weight of the design deviation in the construction stage and the attenuation coefficient of the construction deviation in the operation and maintenance stage includes: inputting the first difference feature as an input variable into the first-level calculation node of the deviation transmission analysis model, wherein the first-level calculation node pre-stores a set of construction process influence factors; weighting and aggregating the first difference feature with each factor in the set of construction process influence factors to obtain the cumulative impact weight of the design deviation in the construction stage; inputting the second difference feature as an input variable into the second-level calculation node of the deviation transmission analysis model, wherein the second-level calculation node pre-stores a set of environmental impact influence factors; weighting and aggregating the second difference feature with each factor in the set of environmental impact influence factors to obtain the initial attenuation coefficient of the construction deviation in the operation and maintenance stage; and feeding back the cumulative impact weight to the second-level calculation node to correct the initial attenuation coefficient to obtain the final attenuation coefficient of the construction deviation in the operation and maintenance stage. This process simulates that construction deviations not only directly affect the operation and maintenance phase, but their magnitude also changes the actual rate at which environmental effects degrade structural performance. By feeding back the cumulative impact weight to correct the attenuation coefficient, the attenuation coefficient can reflect the effect of the deviation magnitude on the performance degradation process, outputting an attenuation characterization that is more in line with the real physical mechanism.

[0011] Preferably, when feeding back the cumulative impact weight to the second-level calculation node, a multiplicative correction method is adopted, in which the cumulative impact weight is used as the multiplier of the attenuation factor to scale the initial attenuation coefficient, so that the correction process has a clear physical meaning, that is, the greater the cumulative design deviation during the construction phase, the more significant the equivalent attenuation of the structure under the same environmental action during the operation and maintenance phase.

[0012] As a preferred technical solution of the present invention, the process of reversely correcting the monitoring parameters of the operation and maintenance phase based on the cumulative influence weight and the attenuation coefficient to generate the corrected operation and maintenance benchmark parameters includes: extracting the current value of the monitoring parameters of the operation and maintenance phase from the multidimensional project data map; multiplying the current value by the cumulative influence weight to obtain the weighted monitoring value; dividing the weighted monitoring value by the attenuation coefficient to obtain the preliminary reverse-inferred equivalent parameters of the construction phase; fusing and calibrating the preliminary reverse-inferred equivalent parameters of the construction phase with the measured parameters of the construction phase stored in the multidimensional project data map; and using the fused and calibrated result as the corrected operation and maintenance benchmark parameters. This reverse correction path utilizes two key coefficients output by the deviation propagation model to reverse-infer the actual construction completion status from the current operation and maintenance monitoring value, and calibrates it in conjunction with actual construction records. This avoids the accumulation of errors that may occur from a single reverse derivation. The obtained operation and maintenance benchmark parameters not only eliminate the cumulative influence of construction deviations but also eliminate the interference of performance degradation during operation and maintenance, and can more realistically reflect the expected state of the engineering structure under conditions of no deviation and no degradation.

[0013] As a preferred embodiment of the present invention, the process of comparing the corrected operation and maintenance baseline parameters with the design parameters from the project initiation stage to determine whether a project design correction instruction needs to be triggered includes: calculating the absolute and relative deviations between the corrected operation and maintenance baseline parameters and the design parameters from the project initiation stage; comparing the absolute deviation with a preset absolute deviation threshold, and simultaneously comparing the relative deviation with a preset relative deviation threshold; when both the absolute deviation and the relative deviation exceed the absolute deviation threshold, generating a project design correction instruction containing the deviation parameter type, and pushing the instruction to the project design end; when neither the absolute deviation nor the relative deviation exceeds the absolute deviation threshold, not generating a project design correction instruction, and marking the current operation and maintenance stage monitoring parameters as acceptable. This judgment logic takes into account both the absolute magnitude and relative proportion of the deviation. Only when both exceed their respective thresholds is it determined that the design scheme has a systematic deviation that needs to be corrected, thereby effectively filtering out acceptable deviations caused by normal construction errors or environmental fluctuations, and avoiding frequent triggering of unnecessary design corrections.

[0014] Preferably, the project design correction instruction containing deviation parameter types is pushed to the project design end, and a comparison curve between the corrected operation and maintenance baseline parameters and the design parameters in the project initiation stage is also pushed, so that designers can intuitively view the change trajectory of deviation parameters in numerical and time dimensions, and assist in making more targeted design adjustment decisions.

[0015] As a preferred embodiment of the present invention, the deviation propagation analysis model adopts a multilayer perceptron network structure, wherein a cross-layer connection channel is set between the first-layer computing nodes and the second-layer computing nodes to directly propagate the cumulative influence weights to the output of the second-layer computing nodes. The existence of the cross-layer connection channel ensures that the cumulative influence weights not only participate in the correction of the attenuation coefficient but also directly affect the final output of the second-layer computing nodes. This provides a shortcut for gradient propagation for the entire deviation propagation network, helps improve the convergence stability during model training, and strengthens the direct impact of construction-stage deviations on the output results of the operation and maintenance stage.

[0016] As a preferred technical solution of the present invention, after generating the corrected operation and maintenance baseline parameters, before comparing the corrected operation and maintenance baseline parameters with the design parameters in the project initiation stage, the following steps are performed: extracting the collection frequency records of the operation and maintenance stage monitoring parameters from the multi-dimensional project data map, calculating the sampling density distribution of the operation and maintenance stage monitoring parameters based on the collection frequency records; identifying sparse sampling periods below a preset density threshold in the sampling density distribution, and extracting the missing parameter types from the sparse sampling periods; for the missing parameter types, calling the values ​​of the same parameter type from the measured parameters in the construction stage for interpolation filling to obtain the completed operation and maintenance stage monitoring parameters, and replacing the original operation and maintenance stage monitoring parameters with the completed operation and maintenance stage monitoring parameters for reverse correction. This step effectively addresses the problem of data sparsity or missing data caused by sensor failures, transmission interruptions, etc. in operation and maintenance monitoring practice, using the measured records in the construction stage to fill the gaps in operation and maintenance data, ensuring the continuity of the input data on which reverse correction depends in time, and avoiding jumps or distortions in the back-calculation results of the baseline parameters due to data missingness.

[0017] As a preferred technical solution of the present invention, after calculating the cumulative impact weight of design deviations during the construction phase and the attenuation coefficient of construction deviations during the operation and maintenance phase, the following steps are further performed: the cumulative impact weights are decomposed according to the time sequence of the construction phases to obtain a sequence of sub-phase impact weights for each construction sub-phase; the attenuation coefficients are decomposed according to the time sequence of the operation and maintenance phases to obtain a sequence of sub-phase attenuation coefficients for each operation and maintenance sub-phase; the sub-phase impact weight sequences and sub-phase attenuation coefficient sequences are cross-compared to identify target sub-phases with abrupt changes in impact weights and a sharp drop in attenuation coefficients, and these target sub-phases are marked as key control nodes throughout the project's entire lifecycle. This approach can precisely pinpoint the critical periods of rapid deviation accumulation and drastic performance degradation from a temporal perspective, providing project managers with a clear window for key control, allowing limited control resources to be prioritized for deployment in the construction or operation and maintenance phases where the deviation amplification effect is strongest.

[0018] As a preferred technical solution of the present invention, after generating the multidimensional project data map, before identifying the first and second difference features, the following steps are performed: The integrity of the design parameters in the project initiation stage, the measured parameters in the construction stage, and the monitoring parameters in the operation and maintenance stage in the multidimensional project data map is verified to identify missing and abnormal parameters; for the identified missing parameters, supplementary values ​​are extracted from the interpolation results of adjacent parameters in the same stage and filled into the missing parameter positions; for the identified abnormal parameters, the normal fluctuation range of the same parameter type under adjacent timestamps is calculated, and the abnormal parameters are corrected to the boundary values ​​of the normal fluctuation range to obtain the cleaned multidimensional project data map. This data cleaning process provides a high-quality input data foundation for subsequent deviation identification and model calculation, avoids the interference of missing and abnormal values ​​on the accuracy of difference feature calculation and the deviation propagation model calculation results, and improves the robustness of the overall analysis process.

[0019] As a preferred technical solution of the present invention, after determining whether a project design correction instruction needs to be triggered, the following steps are further performed: When a project design correction instruction is triggered, upstream and downstream parameter types associated with the deviation parameter type are extracted from the multi-dimensional project data graph to generate a parameter association network; the parameter association network is fused with the cumulative influence weight to calculate the linkage correction coefficient of the upstream and downstream parameter types; the linkage correction coefficient is encapsulated in the project design correction instruction so that the project design end can simultaneously obtain the recommended correction range of the upstream and downstream parameters when receiving the instruction. This optimized solution enables the design end to clearly understand which related parameters will be affected by the adjustment of a certain deviation parameter and the recommended adjustment range when modifying a certain deviation parameter, thereby making more coordinated and systematic design parameter corrections and avoiding secondary problems such as the mismatch of other parameters caused by modifying one parameter.

[0020] This invention also provides a full life-cycle environmental protection project management system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the aforementioned full life-cycle project management method, integrating functions such as parameter acquisition at each stage of the project, data map construction, difference feature identification, deviation transmission analysis, benchmark parameter reverse correction, and design supplementation judgment into a unified processing flow. It can automatically complete the full-link analysis after the monitoring data enters the system, without the need for manual step-by-step operation, thereby improving the automation level and analysis timeliness of the full life-cycle management of the project.

[0021] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0022] By assigning timestamps to design parameters in the project initiation phase, measured parameters in the construction phase, and monitoring parameters in the operation and maintenance phase, and establishing a two-dimensional mapping table with time as the horizontal axis and parameter type as the vertical axis, a multi-dimensional project data map is generated by filling parameters from each phase into the mapping table and associating and labeling the values ​​of the same parameter type under different timestamps. This data map unifies heterogeneous parameters that were originally fragmented across different phases into a single time coordinate system, enabling continuous tracing of the numerical evolution trajectory of any parameter type from the design phase to the construction phase and then to the operation and maintenance phase. When subsequently calculating the first and second difference features, the corresponding values ​​can be directly extracted from the multi-dimensional project data map according to the parameter type to complete the difference encoding and ratio encoding, eliminating the need for cross-system data conversion and time alignment operations, thus simultaneously improving the accuracy and efficiency of difference identification.

[0023] The deviation propagation analysis model adopts a hierarchical computational node structure. The first-layer computational node receives the first difference feature and weights it with a pre-stored set of construction process influencing factors to obtain the cumulative impact weight of the design deviation during the construction phase. The second-layer computational node receives the second difference feature and weights it with a pre-stored set of environmental impact factors to obtain the initial attenuation coefficient of the construction deviation during the operation and maintenance phase. The initial attenuation coefficient is then scaled using the cumulative impact weight as a multiplier of the attenuation factor to obtain the final attenuation coefficient. A cross-layer connection channel is established between the two layers of computational nodes, allowing the cumulative effect of the design deviation to directly participate in the quantitative correction process of the construction deviation attenuation effect. After obtaining the cumulative impact weight and attenuation coefficient, the current value of the monitoring parameters during the operation and maintenance phase is multiplied by the cumulative impact weight and then divided by the attenuation coefficient to obtain the preliminary back-inferred equivalent parameters for the construction phase. These parameters are then further fused and calibrated with the measured parameters of the construction phase stored in the multi-dimensional project data map to generate the corrected operation and maintenance benchmark parameters. This reverse correction process uses the quantitative output of the deviation propagation model to restore the monitoring data during the operation and maintenance phase to a state comparable to the design benchmark. This eliminates the interference of the amplification effect of construction technology and the attenuation effect of environmental effects on deviation judgment. It makes the triggering decision of the design correction instruction based on the direct comparison between the restored benchmark and the original design benchmark, avoiding the design correction being falsely triggered or missed due to ignoring the dynamic characteristics of deviation propagation. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0025] Figure 1It is a flowchart of the whole life cycle engineering project management methodology;

[0026] Figure 2 This is a flowchart of the project design correction and adjustment instruction trigger judgment process;

[0027] Figure 3 This is a flowchart of a method for identifying key control nodes throughout the project lifecycle based on deviation propagation.

[0028] Figure 4 This is a flowchart of multi-dimensional project data map cleaning and linkage correction process;

[0029] Figure 5 It is a time-series comparison chart of beam bottom displacement parameters at multiple stages of a full life cycle engineering project;

[0030] Figure 6 This is a schematic diagram showing the changes in the impact weight of the construction sub-stages and the attenuation coefficient of the operation and maintenance sub-stages throughout the entire life cycle, as well as key control nodes. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] See Figure 1 This invention provides a method for managing environmental protection engineering projects throughout their entire life cycle, comprising the following steps:

[0033] The project obtains design parameters from the project initiation phase, measured parameters from the construction phase, and monitoring parameters from the operation and maintenance phase. The parameters from each phase are then synchronously mapped according to the timeline to generate a multi-dimensional project data map containing timestamps.

[0034] Based on multidimensional project data maps, the first difference feature between design parameters in the project initiation stage and measured parameters in the construction stage is identified, and the second difference feature between measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage is identified.

[0035] The first and second difference features are input into the pre-built deviation transmission analysis model to calculate the cumulative influence weight of the design deviation in the construction stage and the attenuation coefficient of the construction deviation in the operation and maintenance stage.

[0036] Based on the cumulative impact weight and attenuation coefficient, the monitoring parameters in the operation and maintenance phase are reversed to generate the corrected operation and maintenance baseline parameters. The corrected operation and maintenance baseline parameters are then compared with the design parameters in the project initiation phase to determine whether a project design correction instruction needs to be triggered.

[0037] Example 1:

[0038] In practice, the design parameters of the project initiation stage, the measured parameters of the construction stage, and the monitoring parameters of the operation and maintenance stage are obtained. The parameters of each stage are synchronously mapped according to the time axis to generate a multi-dimensional project data map containing timestamps.

[0039] Structural and material design parameters are extracted from the design documents at the project initiation stage. These documents include structural calculation sheets, design drawings, and material lists. Pre-constructed templates for extracting design parameters are used. The structural parameter extraction template includes keywords such as cross-sectional dimensions, span, story height, and reinforcement ratio, while the material parameter extraction template includes keywords such as concrete strength grade, steel reinforcement grade, and modulus of elasticity. The design documents are converted into searchable text using optical character recognition (OCR) or document parsing tools. Regular expressions are then used to match keywords and extract the numerical values ​​and units following the keywords, forming sets of structural and material design parameters.

[0040] Structural and material measurement parameters are extracted from on-site records during the construction phase. These records include construction logs, concealed works acceptance records, material re-inspection reports, and on-site testing records. For structural measurement parameters, the dimensions of completed structural components, protective layer thickness, and weld dimensions are extracted from the construction logs and on-site testing records. For material measurement parameters, the measured compressive strength of concrete test blocks and the measured yield strength of reinforcing steel bars are extracted from the material re-inspection reports. The extraction method involves structuring the tabular data or text descriptions in the on-site records and mapping them to standard parameter names using a predefined parameter dictionary.

[0041] Structural response parameters and environmental impact parameters are extracted from the monitoring system used during the operation and maintenance phase. The monitoring system includes strain sensors, displacement sensors, acceleration sensors, and environmental monitoring stations deployed at key locations on the engineering structure. Structural response parameters, such as strain values, displacement values, and vibration frequencies, as well as environmental impact parameters, such as temperature, humidity, wind load, and vehicle load, are acquired in real time through data acquisition interfaces. The acquired parameters include a timestamp from the acquisition.

[0042] Each extracted parameter is assigned a corresponding timestamp. The design completion time during the project initiation phase is timestamped by the final approval time in the design document approval system. The actual measurement record time during the construction phase is timestamped by the actual measurement date and time recorded in the construction records. The monitoring and acquisition time during the operation and maintenance phase is timestamped by the acquisition time recorded in the sensor data packet. All timestamps are converted to Coordinated Universal Time (UTC) millisecond-level timestamp format to ensure consistency across the entire lifecycle timeline.

[0043] A two-dimensional mapping table is created with time on the horizontal axis and parameter type on the vertical axis. Parameter types include specific parameter names from structural design parameters, material design parameters, structural measured parameters, material measured parameters, structural response parameters, and environmental action parameters, such as "concrete strength" and "beam bottom displacement." Design parameters, measured parameters, and monitoring parameters with timestamps are filled into the corresponding cells of the two-dimensional mapping table. For cases where the same parameter type has multiple values ​​at different timestamps, a numerical sequence for that parameter type is created along the vertical axis in chronological order, and each value in the sequence is labeled with its source stage, forming a multi-dimensional project data map.

[0044] In practice, based on multidimensional project data maps, the first difference between design parameters in the project initiation stage and measured parameters in the construction stage is identified, and the second difference between measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage is identified.

[0045] In the multidimensional project data map, the values ​​of design parameters from the project initiation stage and the measured values ​​of parameters from the construction stage are extracted one by one according to parameter type. For the same parameter type, the design parameter value is denoted as... The measured parameter values ​​during the construction phase are recorded as follows: Calculate the difference between design parameters and measured parameters of the same type. Sum and ratio The difference Sum and ratio Combined encoding into a two-dimensional vector This two-dimensional vector represents the component corresponding to the parameter type in the first differential feature. Performing the same operation on all parameter types yields the first differential feature set.

[0046] Extract the values ​​of measured parameters during the construction phase and monitoring parameters during the operation and maintenance phase according to parameter type. For the same parameter type, the measured parameter value during the construction phase is denoted as... The monitoring parameter values ​​during the operation and maintenance phase are recorded as follows: Calculate the difference between measured parameters and monitored parameters of the same type. Sum and ratio The difference Sum and ratio Combined encoding into a two-dimensional vector This two-dimensional vector represents the component corresponding to the parameter type in the second differential feature. Performing the same operation on all parameter types yields the set of second differential features.

[0047] The first and second difference features are normalized separately to unify their dimensions to a unitless scale. Range normalization is used for the normalization process. For the difference dimension in the first difference feature, the differences corresponding to all parameter types are extracted. Form a set of differences and calculate the minimum value in the set of differences. and maximum value For the ratio dimension in the first difference feature, extract all ratios. Form a set of ratios and calculate the minimum value in the set of ratios. and maximum value The range normalization formula is used.

[0048]

[0049] Differences between parameter types Mapped to The interval is used to obtain the normalized difference. ; Ratios of each parameter type Mapped to The interval is used to obtain the normalized ratio. .in, This represents the original feature values ​​to be normalized, specifically... or ; This represents the minimum value in the corresponding feature set, obtained statistically from the current first set of differential features; This represents the maximum value in the corresponding feature set, obtained statistically from the current first set of differential features. The normalized difference is then calculated. The ratio after normalization Recombined and encoded into a new two-dimensional vector , as the first difference feature after normalization.

[0050] Perform the same range normalization operation on the second difference feature. For the difference dimension in the second difference feature, extract the differences corresponding to all parameter types. Form a set of differences and calculate the minimum value of the set of differences. and maximum value For the ratio dimension, extract all ratios. Form a set of ratios and calculate the minimum value of the set of ratios. and maximum value Using the above range normalization formula, where Replace with and , and The normalized difference is obtained by taking the minimum and maximum values ​​of each set. and normalized ratio The normalized difference The ratio after normalization Recombined and encoded into a new two-dimensional vector This serves as the second differential feature after normalization. After range normalization, the values ​​of both the first and second differential features are within the range... The interval eliminates the influence of different parameter units and numerical ranges.

[0051] See Figure 5 In the graph, the horizontal axis represents time, expressed in relative days to the UTC millisecond timestamp, and the vertical axis represents the beam bottom displacement, expressed in millimeters (mm). The legend shows three types of parameter data: design parameters during the project initiation stage are represented by black dashed lines, measured parameters during the construction stage are marked with hollow squares, and monitoring parameters during the operation and maintenance stage are displayed as black solid curves.

[0052] During the project initiation phase, the design parameter for beam bottom displacement was kept constant at 100mm, represented by a horizontal dashed line, signifying the design baseline value set in the design documents. During the construction phase, the measured beam bottom displacement parameters were distributed over a relative time period of approximately 0 to 180 days, with measured values ​​fluctuating between approximately 98mm and 106mm. Overall, the measured values ​​were higher than the 100mm baseline, reflecting the actual deviation in structural deformation during construction. The construction measurement data points were relatively discrete, reflecting multiple independent measurement records from the field.

[0053] The monitoring parameters during the operation and maintenance phase cover a period of 180 to 600 days, exhibiting continuous curvilinear fluctuations. The overall trend is a gradual decrease from slightly above the design baseline by 100 mm to nearly 95 mm below, demonstrating the dynamic change and gradual reduction of beam bottom displacement over time. The high-frequency oscillations of the monitoring parameters reflect the real-time response characteristics of the structure under the influence of multiple factors such as environment and loads during the operation and maintenance phase.

[0054] This figure visually reflects the temporal evolution of beam bottom displacement parameters throughout the entire lifecycle of an engineering project by synchronously mapping design parameters, measured construction parameters, and operation and maintenance monitoring parameters. The measured parameters during the construction phase significantly deviate from the design baseline, forming the first difference characteristic and providing a basis for the subsequent deviation propagation analysis model. Changes in monitoring parameters during the operation and maintenance phase reflect structural response and environmental impact, representing a significant second difference characteristic. By comparing the measured construction data with the operation and maintenance monitoring data, the changing trends of the structural state and potential deviation attenuation effects can be identified.

[0055] The data in this figure conforms to the construction method of the multidimensional project data map in Example 1, which helps in subsequent normalization processing based on difference features, deviation propagation analysis and reverse correction of operation and maintenance parameters, and provides a quantitative basis for judging the project design correction instructions.

[0056] Example 2:

[0057] In practice, the first and second difference features are input into the pre-built deviation transmission analysis model to calculate the cumulative impact weight of the design deviation in the construction phase and the attenuation coefficient of the construction deviation in the operation and maintenance phase.

[0058] The bias propagation analysis model employs a multilayer perceptron network structure. This structure includes an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer receives normalized first and second differential feature vectors. The first hidden layer corresponds to the first layer of computation nodes in the bias propagation analysis model, containing 256 neurons and using the ReLU activation function. The second hidden layer corresponds to the second layer of computation nodes, containing 128 neurons and using the ReLU activation function. The output layer contains two output neurons, outputting the cumulative influence weight and decay coefficient, respectively. The output neurons use the Sigmoid function to constrain the output values. Interval. A cross-layer connection channel is set between the first-layer computing nodes and the second-layer computing nodes. The cross-layer connection channel directly passes the intermediate feature vector output by the first hidden layer to the output layer, concatenates it with the feature vector output by the second hidden layer, and then inputs it into the fully connected operation of the output layer.

[0059] The training process of the deviation propagation analysis model is as follows: Design parameter samples from the project initiation phase, measured parameter samples from the construction phase, and monitoring parameter samples from the operation and maintenance phase are collected from historical engineering projects. The first and second difference feature samples are extracted and normalized according to the method described in Example 1. These first and second difference feature samples are used as input samples for the deviation propagation analysis model. For each training sample, the weighted average of all parameter type deviations between the measured parameter samples from the construction phase and the design parameter samples from the project initiation phase is used as the true label for the cumulative influence weight, where the weighting coefficient is determined by the magnitude of each factor in the construction process influence factor set. The decay rate of all parameter type deviations between the monitoring parameter samples from the operation and maintenance phase and the measured parameter samples from the construction phase over time is used as the true label for the decay coefficient. The mean squared error loss function is used, and gradient descent training is performed using the Adam optimizer. The learning rate is set to 0.001, the batch size is 32, and the training epochs are 200, resulting in the trained deviation propagation analysis model.

[0060] The set of construction process influencing factors is pre-stored in the parameter storage unit of the first-level calculation node. This set includes influencing factors for formwork engineering, concrete pouring, curing conditions, and construction loads. The formwork engineering influencing factor ranges from 0.8 to 1.2, determined by the compliance rate of formwork type, formwork joint tightness, and formwork removal time from the construction records. A value of 1.0 is used when the compliance rate meets national standards; a value between 0.8 and 1.0 is used when the compliance rate is lower than national standards; and a value between 1.0 and 1.2 is used when the compliance rate is higher than national standards. The concrete pouring influencing factor ranges from 0.7 to 1.3, determined by the degree of compliance of the vibration method, layer thickness, and pouring interval time during the pouring process with the construction specifications. A value of 1.0 is used when the construction specifications are fully met; a value less than 1.0 is used when vibration is insufficient or missed; and a value greater than 1.0 is used when vibration is excessive. The value of the curing condition influence factor ranges from 0.6 to 1.4, determined based on the deviations of the curing method, curing time, and temperature and humidity from the design requirements. A value of 1.0 is taken when the design requirements are met, less than 1.0 when curing is insufficient, and greater than 1.0 when curing is intensified. The value of the construction load influence factor ranges from 0.9 to 1.1, determined based on the ratio of the actual applied load during construction to the design allowable construction load. A value of 1.0 is taken when the ratio equals the design allowable value.

[0061] The set of environmental impact factors is pre-stored in the parameter storage unit of the second-level calculation node. This set includes temperature impact factor, humidity impact factor, load cycle factor, and medium erosion factor. The temperature impact factor ranges from 0.85 to 1.15, determined by the ratio of the annual average temperature during the operation and maintenance phase to the design reference temperature. A value of 1.0 is used when the ratio equals 1, greater than 1.0 when the ratio is greater than 1, and less than 1.0 when the ratio is less than 1. The humidity impact factor ranges from 0.80 to 1.20, determined by the ratio of the annual average relative humidity during the operation and maintenance phase to the design reference humidity. A value of 1.0 is used when the ratio equals 1. The load cycle factor ranges from 0.90 to 1.10, determined by the ratio of the actual cumulative fatigue load cycles to the design fatigue life. A value of 1.0 is used when the ratio equals 1. The value of the medium corrosion factor ranges from 0.75 to 1.25, and is determined based on the ratio of the concentration of the environmental corrosive medium to the allowable concentration corresponding to the design protection level. When the ratio is equal to 1, it is taken as 1.0.

[0062] During the model inference phase, the normalized first difference feature is input as an input variable to the first-layer computation node of the deviation propagation analysis model. The first-layer computation node multiplies the first difference feature vector with the weight matrix of the first hidden layer and adds a bias vector, then passes it through the ReLU activation function to obtain the first intermediate feature vector. The first-layer computation node reads the current values ​​of the formwork engineering influence factor, concrete pouring influence factor, curing condition influence factor, and construction load influence factor from the construction process influence factor set in the parameter storage unit, constructing a four-dimensional influence factor vector from the four influence factors. An element-wise weighted aggregation operation is performed on the first intermediate feature vector and the four-dimensional influence factor vector, multiplying each element of the first intermediate feature vector with a corresponding element of the four-dimensional influence factor vector and summing the results to obtain the cumulative influence weight of the design deviation during the construction phase. The numerical value of the cumulative influence weight represents the comprehensive degree to which the deviation between the design parameters in the project initiation phase and the measured parameters in the construction phase is amplified or reduced under the combined effect of various stages of the construction process. A value greater than 1.0 indicates that the deviation is cumulatively amplified, and a value less than 1.0 indicates that the deviation is partially reduced.

[0063] The normalized second difference feature is input as an input variable to the second-layer computation node of the deviation propagation analysis model. The second-layer computation node multiplies the second difference feature vector with the weight matrix of the second hidden layer and adds a bias vector, then passes it through the ReLU activation function to obtain the second intermediate feature vector. The second-layer computation node reads the current values ​​of the temperature factor, humidity factor, load cycle factor, and medium erosion factor from the environmental influence factor set in the parameter storage unit, constructing a four-dimensional influence factor vector from these four factors. An element-wise weighted aggregation operation is performed on the second intermediate feature vector and the four-dimensional influence factor vector, multiplying each element of the second intermediate feature vector with a corresponding element of the four-dimensional influence factor vector and summing the results to obtain the initial attenuation coefficient of the construction deviation during the operation and maintenance phase. The initial attenuation coefficient represents the proportion of deviation between the construction and operation and maintenance phases that attenuates over time under the influence of various environmental factors. The range is such that a value close to 0 indicates significant decay, while a value close to 1 indicates virtually no decay.

[0064] The cumulative impact weight is fed back to the second-layer calculation node to correct the initial attenuation coefficient, resulting in the final attenuation coefficient of the construction deviation during the operation and maintenance phase. A multiplicative correction method is used during feedback, with the cross-layer connection channel directly sending the cumulative impact weight as an independent feature component to the output layer. In the output layer, the cumulative impact weight, the initial attenuation coefficient, and the intermediate features transmitted through the cross-layer connection channel are concatenated, and the final output is obtained through a fully connected operation. The multiplicative correction method uses the cumulative impact weight as the multiplier of the attenuation factor, and the final attenuation coefficient is calculated as follows:

[0065]

[0066] in, This represents the final attenuation coefficient, and its value is within... interval; This represents the initial attenuation coefficient, which is output by the second-layer calculation node after weighted aggregation of the set of second differential characteristics and environmental influence factors. This represents the cumulative impact weight, which is output by the first-level calculation node after weighted aggregation of the first difference feature and the set of construction technology impact factors. This represents the learnable weight parameters assigned to the decay path in the output layer, which are updated through backpropagation during the training of the bias propagation analysis model. This represents the bias term of this computation path in the output layer, which is updated through backpropagation during the training of the bias propagation analysis model. This represents the Sigmoid activation function, which maps the product result to... Interval.

[0067] In practice, the monitoring parameters during the operation and maintenance phase are reverse-corrected based on the cumulative impact weight and attenuation coefficient to generate the corrected operation and maintenance baseline parameters.

[0068] The current values ​​of the operation and maintenance (O&M) phase monitoring parameters are extracted from the multidimensional project data map. These current values ​​are the collected values ​​obtained within the most recent sampling period, including structural response parameters and environmental impact parameters.

[0069] The current value of the monitoring parameter during the operation and maintenance phase is multiplied by the cumulative impact weight to obtain the weighted monitoring value. For each parameter type, the weighted monitoring value is equal to the product of the current value of the monitoring parameter during the operation and maintenance phase for that parameter type and the cumulative impact weight. When the cumulative impact weight is greater than 1.0, the weighted monitoring value is greater than the original current value of the monitoring parameter during the operation and maintenance phase; when the cumulative impact weight is less than 1.0, the weighted monitoring value is less than the original current value of the monitoring parameter during the operation and maintenance phase.

[0070] Dividing the weighted monitoring values ​​by the final attenuation coefficient yields the preliminary back-calculated equivalent parameters for the construction phase. These preliminary back-calculated equivalent parameters are calculated by inversely estimating the monitoring parameters for the operation and maintenance phase after they have undergone the combined effects of accumulated construction deviations and attenuation during the operation and maintenance phase. The calculation process involves multiplying the current value of the monitoring parameter for the operation and maintenance phase by the cumulative influence weight and then dividing by the final attenuation coefficient.

[0071] The preliminary inverse calculation of equivalent parameters for the construction stage is fused and calibrated with the measured parameters for the construction stage stored in the multidimensional project data map. The fusion calibration uses a weighted average method, with the weighting coefficients determined by the acquisition confidence level of the measured parameters for the construction stage and the calculated confidence level of the preliminary inverse calculation of the equivalent parameters. The acquisition confidence level of the measured parameters for the construction stage is determined based on the accuracy level of the testing method: 0.9 for destructive testing and 0.7 for non-destructive testing; the calculated confidence level of the preliminary inverse calculation of the equivalent parameters is 0.6. The measured parameters for the construction stage are multiplied by the acquisition confidence level, and this is combined with the preliminary inverse calculation of the equivalent parameters multiplied by the calculated confidence level. The sum of these two results is divided by the sum of the acquisition confidence level and the calculated confidence level to obtain the fused calibration result. This fused calibration result is used as the corrected operation and maintenance baseline parameters. The corrected operation and maintenance baseline parameters are used to replace the directly acquired operation and maintenance stage monitoring parameters as a benchmark for judging whether the engineering structure deviates from the design intent.

[0072] Example 3:

[0073] In specific implementation, please refer to Figure 2 The revised operation and maintenance baseline parameters are compared with the design parameters from the project initiation stage to determine whether a project design correction instruction needs to be triggered.

[0074] The absolute and relative deviations between the corrected operation and maintenance baseline parameters and the design parameters from the project initiation phase are calculated. The corrected operation and maintenance baseline parameters are extracted from the multi-dimensional project data map and consist of a set of values ​​for various parameter types. The values ​​for each parameter type have undergone inverse correction and fusion calibration processing using cumulative influence weights and attenuation coefficients. The design parameters from the project initiation phase are extracted from the multi-dimensional project data map and consist of the original design values ​​for the same parameter types as the corrected operation and maintenance baseline parameters. For each parameter type, the values ​​for that parameter type in the corrected operation and maintenance baseline parameters and the values ​​in the project initiation phase design parameters are extracted separately. The absolute deviation is calculated as the absolute value of the difference between the corrected operation and maintenance baseline parameter values ​​and the project initiation phase design parameter values. The relative deviation is calculated by dividing the absolute deviation by the absolute value of the project initiation phase design parameter value. When the project initiation phase design parameter value is zero, the relative deviation is directly taken as the absolute deviation value.

[0075] In a specific calculation process, for the first... Parameter type, absolute deviation The calculation formula is:

[0076]

[0077] in, Indicates the first The absolute deviation corresponding to each parameter type; This indicates the first parameter in the revised operation and maintenance baseline parameters. A numerical value of a parameter type, which is read from the output of the fusion calibration step; Indicates the first design parameter in the project initiation stage A numerical value of type parameter, read from the design parameter set of the project initiation phase of the multidimensional project data map. Relative deviation. pass Calculations show that when When equal to zero, Take directly The value of .

[0078] The system compares the absolute deviation with a preset absolute deviation threshold, and the relative deviation with a preset relative deviation threshold. The preset absolute and relative deviation thresholds are pre-stored in the system configuration database. The preset absolute deviation threshold is a set of values ​​corresponding one-to-one with each parameter type, obtained from engineering structural design codes and construction quality acceptance standards. For structural dimension parameters, the preset absolute deviation threshold is set according to the allowable deviation values ​​specified in the "Code for Acceptance of Construction Quality of Concrete Structures" GB50204. For example, the preset absolute deviation threshold for beam cross-section dimensions is 8mm, and for column verticality parameters it is 10mm. For material strength parameters, the preset absolute deviation threshold is set according to the lower limit value corresponding to the design strength grade. The preset relative deviation threshold is also set separately for each parameter type, uniformly adopting the acceptable deviation percentage specified in the general engineering design instructions. For example, the preset relative deviation threshold for material strength parameters is set to 5%, and for structural natural frequency parameters it is set to 3%. The absolute deviation for each parameter type... The relative deviation is compared with the corresponding preset absolute deviation threshold. Compare with the corresponding preset relative deviation threshold.

[0079] When the absolute deviation of any parameter type exceeds the corresponding preset absolute deviation threshold, and the relative deviation of the same parameter type also exceeds the corresponding preset relative deviation threshold, a project design correction instruction containing the deviation parameter type is generated. The project design correction instruction is a structured message. The message body includes an instruction type field, a timestamp field, a list of parameter types exceeding limits field, and a deviation details field. The instruction type field has the value "Design Correction". The list of parameter types exceeding limits field records the names of all parameter types that simultaneously meet the conditions of exceeding both absolute and relative deviation limits. The deviation details field is a nested array, with each array element corresponding to an parameter type exceeding limits, containing the corrected operation and maintenance baseline parameter value, the design parameter value from the project initiation stage, the absolute deviation value, the relative deviation value, the corresponding preset absolute deviation threshold, and the preset relative deviation threshold. The project design correction instruction is pushed to the message receiving queue on the project design side through the system's internal message middleware. The project design side subscribes to the queue and parses the instruction content for display and notification.

[0080] When the absolute deviation of all parameter types does not exceed the corresponding preset absolute deviation threshold, or the relative deviation of all parameter types does not exceed the corresponding preset relative deviation threshold, no project design correction instruction will be generated. The system will mark the collected values ​​corresponding to each parameter type in the current operation and maintenance phase monitoring parameters as acceptable. The marking operation for acceptable status is to add a status attribute field to each parameter type record of the operation and maintenance phase monitoring parameters at the current timestamp in the multi-dimensional project data graph. The key of the status attribute field is "acceptability", and the value is "acceptable", indicating that the actual status of the engineering structure within the current monitoring period is within the design-allowed deviation range, and no design correction is required.

[0081] When a project design correction instruction containing deviation parameter types is pushed to the project design end, a comparison curve between the corrected operation and maintenance baseline parameters and the design parameters from the project initiation phase is also pushed. The comparison curve is plotted with time on the horizontal axis and parameter values ​​on the vertical axis, generating a separate comparison curve for each parameter type that simultaneously exceeds limits. In the comparison curve, the sequence of changes in the corrected operation and maintenance baseline parameters over time is plotted as a solid curve, with data points derived from the corrected operation and maintenance baseline parameter values ​​for that parameter type at various timestamps in the multi-dimensional project data graph; the design parameters from the project initiation phase are plotted as a horizontal dashed line, with the vertical axis value being the design parameter value from the project initiation phase for that parameter type. On the comparison curve, vertical lines or shaded areas are used to mark the time intervals where both the absolute deviation and relative deviation exceed a preset absolute deviation threshold and a preset relative deviation threshold. The comparison curve is encoded as a binary data stream in PNG format and packaged together with the project design correction instruction for push via message middleware.

[0082] Example 4:

[0083] In specific implementation, please refer to Figure 3 After generating the corrected operation and maintenance baseline parameters, before comparing the corrected operation and maintenance baseline parameters with the design parameters in the project initiation stage, a step of sparse sampling and completion of the monitoring parameters in the operation and maintenance stage is also performed.

[0084] The sampling frequency records of monitoring parameters during the operation and maintenance phase are extracted from the multidimensional project data map. These records are stored in the metadata layer of the multidimensional project data map. Each sampling frequency record is associated with a parameter type and a timestamp range. The record contains a list of sampling times and a sequence of time intervals between adjacent sampling times. The sampling frequency records are scanned for each parameter type, extracting the time interval values ​​between each adjacent sampling time throughout the entire lifecycle of the operation and maintenance phase. These time interval values ​​are stored in seconds.

[0085] The sampling density distribution of monitoring parameters during the operation and maintenance phase is calculated based on the collection frequency records. The function for constructing the sampling density distribution is as follows: taking the entire time interval of the operation and maintenance phase as the domain, the time interval is uniformly divided into several time windows of fixed length and width, with the time window width set to 86400 seconds. For the first... Within a given time window, the number of samples taken within that time window is counted. This number of samples is then divided by the time window width to obtain the nth sample. The average sampling density within each time window is expressed in Hertz. By iterating through all time windows, a one-dimensional array of the sampling density over time is obtained for each parameter type. The index of the one-dimensional array corresponds to the sequence number of the time window, and the element value of the one-dimensional array is the average sampling density. This one-dimensional array is the sampling density distribution for that parameter type.

[0086] Identify sparse sampling periods in the sampling density distribution that are below a preset density threshold. The preset density threshold is stored in the system configuration parameter table. The preset density threshold is set based on the following criteria: For structural response parameters, the preset density threshold is determined according to the minimum sampling frequency specified in the structural health monitoring standard, which is 0.0001 Hz (meaning at least one sample is collected every 10,000 seconds). For environmental impact parameters, the preset density threshold is determined according to the minimum sampling frequency specified in the environmental monitoring specification, which is 0.00001157 Hz (meaning at least one sample is collected every 86,400 seconds). The average sampling density of each time window in the sampling density distribution is compared with the corresponding preset density threshold. Time windows with an average sampling density below the preset density threshold are marked as sparse sampling windows. Consecutive sparse sampling windows on the time axis are merged into a continuous time interval, which is considered a sparse sampling period.

[0087] Extract missing parameter types from sparse sampling periods. For each sparse sampling period, query the stored records of monitoring parameters during the operation and maintenance phase within that time interval in the multidimensional project data map. When the sampling count of a certain parameter type is zero within that time interval, mark that parameter type as a missing parameter type. The record for missing parameter types includes the start timestamp, end timestamp, and name of the missing parameter type for the sparse sampling period.

[0088] For missing parameter types, interpolation is performed using values ​​of the same parameter type from the measured parameters during the construction phase. In the set of measured parameters during the construction phase, parameter types with the same name as the missing parameter type are retrieved, and the measured value sequences for that parameter type under all timestamps within the construction phase are extracted. The extracted measured value sequences are arranged in ascending order of timestamps to form the data source sequence to be filled. For each missing sampling moment within the sparse sampling period, the nearest neighbor interpolation method is used to select values ​​from the data source sequence to be filled. The nearest neighbor interpolation method selects the measured value of the construction phase that is closest to the missing sampling moment in time distance, measured by the absolute time difference between the missing sampling moment and the actual construction phase recording moment. The selected measured value of the construction phase is used as the filling value for that missing sampling moment, and a data source marker "Construction Phase Interpolation Fill" is added to obtain the completed operation and maintenance phase monitoring parameters. The completed operation and maintenance phase monitoring parameters replace the original operation and maintenance phase monitoring parameters in the multidimensional project data map, and are used for extracting the current values ​​of the operation and maintenance phase monitoring parameters in subsequent reverse correction steps.

[0089] In practice, after calculating the cumulative impact weight of design deviations during the construction phase and the attenuation coefficient of construction deviations during the operation and maintenance phase, the step of identifying key control nodes is also performed.

[0090] The cumulative impact weights are decomposed according to the chronological order of the construction stages to obtain the sub-stage impact weight sequence for each construction sub-stage. The start and end times of the construction stages are determined from the first and last timestamps of the measured parameters of the construction stage in the multi-dimensional project data map. The entire time interval of the construction stage is divided into several construction sub-stages according to the construction procedures. The division of construction sub-stages is based on the sub-project or sub-item project division scheme in the construction organization design document of the project. The construction sub-stages include the foundation construction sub-stage, the main structure construction sub-stage, the roof construction sub-stage, and the decoration and finishing construction sub-stage. Within the time interval of each construction sub-stage, the intermediate cumulative impact weight values ​​of each deviation calculation output within that sub-stage are obtained from the first-level calculation node of the deviation propagation analysis model. The arithmetic mean of the intermediate cumulative impact weight values ​​within the same construction sub-stage is taken to obtain the sub-stage impact weight of that construction sub-stage. The sub-stage impact weights are arranged in chronological order of the construction sub-stages to form the sub-stage impact weight sequence.

[0091] The attenuation coefficients are decomposed according to the time sequence of the operation and maintenance (O&M) phases to obtain a sub-phase attenuation coefficient sequence for each O&M sub-phase. The start and end times of the O&M phases are determined by the first timestamp of the O&M phase monitoring parameters in the multi-dimensional project data map and the current time. The entire time interval of the O&M phase is divided into several O&M sub-phases at fixed time intervals, with the fixed time interval set to 2,629,800 seconds, i.e., the number of seconds in one month. Within the time interval of each O&M sub-phase, the final attenuation coefficient value of each deviation transmission calculation output within that sub-phase is obtained from the output layer of the deviation transmission analysis model. The arithmetic mean of the final attenuation coefficient values ​​within the same O&M sub-phase is taken to obtain the sub-phase attenuation coefficient for that O&M sub-phase. The sub-phase attenuation coefficients are arranged in chronological order of the O&M sub-phases to form a sub-phase attenuation coefficient sequence.

[0092] The sub-stage impact weight sequence and the sub-stage decay coefficient sequence are cross-compared. The cross-comparison method is as follows: each sub-stage impact weight in the sub-stage impact weight sequence is paired with the corresponding time-based decay coefficient of the maintenance sub-stage in the sub-stage decay coefficient sequence. The pairing rule is to align the end time of the construction sub-stage with the start time of the maintenance sub-stage. The construction sub-stage with the highest time overlap after alignment forms a comparison pair with the maintenance sub-stage. For each comparison pair, the rate of change of the sub-stage impact weight and the rate of change of the sub-stage decay coefficient are calculated. The rate of change of the sub-stage impact weight is obtained by subtracting the sub-stage impact weight of the previous construction sub-stage from the current construction sub-stage's sub-stage impact weight, and then dividing by the previous construction sub-stage's sub-stage impact weight. The rate of change of the sub-stage decay coefficient is obtained by subtracting the sub-stage decay coefficient of the previous maintenance sub-stage from the current maintenance sub-stage's sub-stage decay coefficient, and then dividing by the previous maintenance sub-stage's sub-stage decay coefficient.

[0093] Identify the target sub-stages where the impact weights change abruptly and the attenuation coefficient drops sharply. The criteria for a sudden change in impact weights are: the rate of change of the sub-stage's impact weights exceeds a preset threshold of 0.3, based on the fact that when construction techniques or load conditions change significantly, the cumulative impact weight jump typically exceeds 30%. The criteria for a sharp drop in the attenuation coefficient are: the rate of change of the sub-stage's attenuation coefficient is negative and its absolute value exceeds a preset threshold of 0.2, based on the fact that when environmental effects suddenly intensify or the structure enters an accelerated deterioration phase, the decrease in the attenuation coefficient typically exceeds 20%. When a pair of comparison pairs simultaneously meets the conditions where the rate of change of the sub-stage's impact weights exceeds the weight change rate threshold, and the rate of change of the sub-stage's attenuation coefficient is negative and its absolute value exceeds the attenuation change rate threshold, the time intersection region of the construction sub-stage and the operation and maintenance sub-stage corresponding to that comparison pair is identified as the target sub-stage. Mark the target sub-phase as a key control node in the entire project lifecycle. The marking operation adds a special event marker to the time axis of the multidimensional project data graph. The type field value of the special event marker is "key control node". The attribute fields include the sub-phase impact weight value of the construction sub-phase, the sub-phase attenuation coefficient value of the operation and maintenance sub-phase, the rate of change of the sub-phase impact weight, and the rate of change of the sub-phase attenuation coefficient.

[0094] See Figure 6In the graph, the horizontal axis represents the step number of the entire project lifecycle, with the first half representing the construction sub-stage number and the second half representing the operation and maintenance sub-stage number. The left side of the vertical axis corresponds to the influence weight of the sub-stage, and the right side corresponds to the attenuation coefficient of the sub-stage. The solid curve in the graph represents the change in the influence weight of the construction sub-stage, and the dashed curve represents the change in the attenuation coefficient of the operation and maintenance sub-stage. The gray shaded areas in the graph are marked as key control nodes, reflecting the critical time intervals of the construction and operation and maintenance sub-stages.

[0095] As shown in the figure, the influence weight of each construction sub-stage exhibits a step-like trend. Initially, the influence weight is approximately 0.95. Subsequently, two significant upward steps occur between steps 200 and 400, reaching a maximum of 1.20, indicating that design deviations are amplified by the cumulative influence of construction technology factors during the construction phase. The influence weight then decreases in subsequent construction sub-stages, eventually approaching 1.00, showing that the cumulative effect of deviations has weakened.

[0096] The attenuation coefficient of the operation and maintenance sub-phase gradually decreases from step 600, exhibiting a clear exponential decay trend, decreasing from approximately 0.83 to nearly 0.35, indicating that the construction deviation is gradually reduced by environmental influences during the operation and maintenance phase. Two distinct fluctuating decreasing intervals exist in the attenuation curve, corresponding to the gray critical control node time periods, located at approximately steps 600 to 700 and 750 to 850 respectively, indicating that the attenuation effect of environmental factors on structural deviations is significantly enhanced during these time intervals.

[0097] The step change in the sub-stage influence weights and the continuous decrease in the sub-stage attenuation coefficients complement each other, and both show significant changes at key control nodes. This aligns with the technical solution described in Example 4, which identifies key control nodes through time alignment and rate of change analysis of the sub-stage influence weight sequence and the sub-stage attenuation coefficient sequence. The figure reflects the cumulative amplification effect of design deviations caused by construction process influencing factors during the construction phase and their gradual attenuation process under the influence of environmental factors during the operation and maintenance phase. It also clearly identifies the key time intervals for focused control of the project throughout its entire lifecycle.

[0098] Example 5:

[0099] In specific implementation, please refer to Figure 4 After generating the multidimensional project data map, before identifying the first and second difference features, a data cleaning step is performed on the multidimensional project data map.

[0100] Completeness verification was performed on the design parameters during the project initiation phase, the measured parameters during the construction phase, and the monitoring parameters during the operation and maintenance phase in the multidimensional project data map. The completeness verification process was as follows: In the two-dimensional mapping table of the multidimensional project data map, each parameter type was scanned row by row. For each parameter type, all timestamp positions of that parameter type on the entire lifecycle timeline were extracted. The timeline was segmented according to the project initiation phase, construction phase, and operation and maintenance phase. Within each phase's time interval, it was checked whether at least one valid numerical record existed for that parameter type. If no numerical record existed for a parameter type within the project initiation phase time interval, the corresponding position of that parameter type in the project initiation phase was marked as a missing parameter. Similarly, if no numerical record existed for a parameter type within the construction phase time interval, the corresponding position of that parameter type in the operation and maintenance phase was marked as a missing parameter. Simultaneously, outlier detection is performed on the numerical records within each stage. Outlier detection employs a statistical threshold discrimination method: for all numerical values ​​of the same parameter type within the same stage, the mean of the numerical set is calculated. and standard deviation subscript Indicates the stage identifier, with a value of These correspond to the project initiation phase, construction phase, and operation and maintenance phase, respectively. A value is considered satisfactory when its deviation from the mean of that phase exceeds three times the standard deviation. When this value is displayed, it is marked as an abnormal parameter. This indicates a specific numerical value for this parameter type within the corresponding stage. This indicates that the parameter type is in the stage. The arithmetic mean of all values ​​in the set. This indicates that the parameter type is in the stage. The standard deviation of all values ​​within the range. The standard deviation is calculated as follows: ,in Representation phase The total number of values ​​of this parameter type. Representation phase Inner A number.

[0101] For identified missing parameters, supplementary values ​​are extracted from the interpolation results of adjacent parameters in the same stage and filled into the missing parameter's position. Adjacent parameters in the same stage are determined as follows: in the two-dimensional mapping table of the multidimensional project data map, based on the parameter type of the missing parameter, adjacent parameter types that are physically related to the missing parameter type within the same time interval are searched. The criteria for determining adjacent parameter types are the mechanical transmission relationship or spatial adjacency relationship between the parameters of various components in the engineering project structure. For example, for the mid-span deflection parameter type of a beam, adjacent parameter types are the support displacement parameter type and the mid-span section strain parameter type of the same beam component. After determining the adjacent parameter types, the values ​​of the adjacent parameter types at the timestamp corresponding to the missing parameter are extracted. A linear interpolation method is used, with the values ​​of the adjacent parameter types as interpolation nodes, to calculate the supplementary values ​​for the missing parameters. The interpolation weights are determined based on the correlation coefficient between the adjacent parameter types and the missing parameter type, which is extracted from the sensitivity matrix in the engineering project structural mechanics model. The calculated supplementary values ​​are filled into the corresponding positions of the missing parameters in the multidimensional project data map, and a data source marker "same-stage adjacent parameter interpolation supplement" is added to these values.

[0102] For identified anomalous parameters, calculate the normal fluctuation range of the same parameter type at adjacent timestamps. The normal fluctuation range is calculated as follows: in the multidimensional project data map, extract the value of the same parameter type as the anomalous parameter at a timestamp preceding the time corresponding to the anomalous parameter. and the value at a timestamp following the time corresponding to the abnormal parameter. .by and Based on this, the normal fluctuation range is ,in , , This is the fluctuation tolerance coefficient. The value is 0.1. The setting of 0.1 is based on the fact that in the field of engineering monitoring, the normal fluctuation range between adjacent sampling points usually does not exceed 10% of the numerical change between adjacent sampling points. Abnormal parameters are corrected to the boundary value of the normal fluctuation range: when the abnormal parameter value is greater than... When the abnormal parameter value is replaced with When the abnormal parameter value is less than When the abnormal parameter value is replaced with Add a data source marker, "Outlier Boundary Correction," to the replaced values. After completing all missing parameter imputation and outlier parameter correction, the cleaned multidimensional project data map is obtained.

[0103] In practice, after determining whether a project design correction instruction needs to be triggered, the steps of calculating and pushing the linkage correction coefficient must also be performed.

[0104] When a project design correction instruction is triggered, upstream and downstream parameter types associated with the deviation parameter type are extracted from the multidimensional project data graph to generate a parameter association network. The deviation parameter type refers to the parameter type recorded in the project design correction instruction that simultaneously satisfies both absolute and relative deviation exceedance conditions. For each deviation parameter type, other parameter types with direct mechanical transfer relationships or design constraint relationships with the deviation parameter type are retrieved from the parameter type relationship graph of the multidimensional project data graph. Direct mechanical transfer relationships are derived from the finite element stiffness matrix of the engineering project structure. A direct mechanical transfer relationship is determined when the stiffness coefficient between the degree of freedom corresponding to the deviation parameter type and the degree of freedom corresponding to other parameter types in the finite element stiffness matrix is ​​non-zero. Design constraint relationships are obtained from the design formulas and construction requirements extracted from the engineering project design documents. A design constraint relationship is determined when two parameter types simultaneously appear on both sides of the equal sign of the same design formula or in the upper and lower limit conditions of the same construction requirement. Among parameter types that have a direct mechanical transfer relationship or design constraint relationship with the deviation parameter type, the parameter type upstream of the deviation parameter type in the physical causal chain is marked as the upstream parameter type, and the parameter type downstream of the deviation parameter type is marked as the downstream parameter type. Using the deviation parameter type as the central node and all associated upstream and downstream parameter types as adjacent nodes, they are connected by directed edges, with the edges pointing from the upstream parameter type to the deviation parameter type and from the deviation parameter type to the downstream parameter type, thus constructing a parameter association network.

[0105] The parameter correlation network and cumulative impact weights are fused to calculate the linkage correction coefficients for upstream and downstream parameter types. The fusion calculation method is as follows: the cumulative impact weights are extracted from the output of the deviation propagation analysis model. These cumulative impact weights are scalar values ​​obtained by weighted aggregation of design deviations during the construction phase using a set of construction process influence factors. For each upstream and downstream parameter type in the parameter correlation network, the normalized result of the first difference feature for that parameter type is extracted from the multidimensional project data map. The normalized difference value in the normalized result of the first difference feature is then used to calculate the correlation correction coefficients for upstream and downstream parameter types. As an influencing factor, the linkage correction coefficient. The calculation formula is: .in, Indicates the parameter association network of the first Linkage correction coefficients for upstream or downstream parameter types The value indicates the recommended correction percentage for this parameter type when the deviation parameter type deviates. The range of values ​​for is real numbers greater than 0. A value greater than 1.0 indicates that the correction magnitude for this parameter type needs to be greater than the correction magnitude of the deviation parameter type itself. A value less than 1.0 indicates that the correction magnitude can be smaller than the correction magnitude of the deviation parameter type itself; Indicates the first The normalized difference in the first difference feature of each parameter type is calculated from the cleaned multidimensional project data map. The range of values ​​is ; This represents the cumulative impact weight of design deviations during the construction phase, output by the first-level calculation node of the deviation propagation analysis model. The range of values ​​for is real numbers greater than 0; Indicates the first The edge weight coefficients between each parameter type and the deviation parameter type in the parameter association network are obtained by normalizing the absolute values ​​of the stiffness coefficients between the corresponding degrees of freedom in the finite element stiffness matrix of the engineering project structure. The range of values ​​is When the first When the absolute value of the stiffness coefficient between a parameter type and a deviation parameter type is zero, the edge weight coefficient does not exist, and this parameter type is not included in the parameter association network. Calculate the linkage correction coefficients corresponding to each upstream and downstream parameter type in the parameter association network to form a set of linkage correction coefficients.

[0106] The set of linkage correction coefficients is encapsulated within the project design backfill correction instruction. The encapsulation method is as follows: In the deviation details field of the project design backfill correction instruction message body, a "Recommended Related Parameter Correction" subfield is added for each nested object corresponding to each deviation parameter type. The value of the "Recommended Related Parameter Correction" subfield is an array, where each element contains the name of the related parameter type, the upstream and downstream attribute identifiers to which the related parameter type belongs, and the linkage correction coefficient value. The encapsulated project design backfill correction instruction is pushed to the message receiving queue of the project design end through a message middleware. When the project design end receives the project design backfill correction instruction, it simultaneously displays the over-limit details of the deviation parameter type and the recommended correction range for all related upstream and downstream parameter types on the display interface.

[0107] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for managing environmental protection engineering projects throughout their entire life cycle, characterized in that, Includes the following steps: Obtain design parameters from the project initiation phase, measured parameters from the construction phase, and monitoring parameters from the operation and maintenance phase. Synchronously map the parameters of each phase according to the time axis to generate a multi-dimensional project data map containing timestamps. Based on multidimensional project data maps, the first difference feature between design parameters in the project initiation stage and measured parameters in the construction stage is identified, and the second difference feature between measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage is identified. The first and second difference features are input into the pre-built deviation transmission analysis model to calculate the cumulative impact weight of the design deviation in the construction stage and the attenuation coefficient of the construction deviation in the operation and maintenance stage. Based on the cumulative impact weight and attenuation coefficient, the monitoring parameters in the operation and maintenance phase are reversed to generate the corrected operation and maintenance baseline parameters. The corrected operation and maintenance baseline parameters are then compared with the design parameters in the project initiation phase to determine whether it is necessary to trigger the project design backfill correction instruction. The specific steps for obtaining design parameters from the project initiation phase, measured parameters from the construction phase, and monitoring parameters from the operation and maintenance phase, and synchronously mapping the parameters of each phase according to the time axis to generate a multi-dimensional project data map containing timestamps are as follows: Structural design parameters and material design parameters are extracted from the design documents during the project initiation stage, structural measured parameters and material measured parameters are extracted from the on-site records during the construction stage, and structural response parameters and environmental action parameters are extracted from the monitoring system during the operation and maintenance stage. Each extracted parameter is assigned a corresponding timestamp, which records the design completion time during the project initiation phase, the actual measurement record time during the construction phase, and the monitoring and data collection time during the operation and maintenance phase. A two-dimensional mapping table is established with time as the horizontal axis and parameter type as the vertical axis. Design parameters, measured parameters, and monitoring parameters with timestamps are filled into the two-dimensional mapping table, and the values ​​of the same parameter type under different timestamps are associated and labeled to form a multi-dimensional project data map.

2. The method for managing environmental protection engineering projects throughout their entire life cycle according to claim 1, characterized in that, The specific steps for identifying the first difference feature between design parameters in the project initiation stage and measured parameters in the construction stage, and the second difference feature between measured parameters in the construction stage and monitoring parameters in the operation and maintenance stage, based on multi-dimensional project data maps, are as follows: In the multidimensional project data map, the values ​​of design parameters in the project initiation stage and measured parameters in the construction stage are extracted one by one according to parameter type. The difference and ratio between the same type of design parameters and measured parameters are calculated, and the difference and ratio are combined and encoded as the first difference feature. According to the parameter type, extract the values ​​of the measured parameters in the construction stage and the values ​​of the monitoring parameters in the operation and maintenance stage one by one, calculate the difference and ratio between the measured parameters and the monitoring parameters of the same type, and encode the difference and ratio as the second difference feature. The first and second difference features are normalized respectively to unify their dimensions to a unitless scale.

3. The method for managing environmental protection engineering projects throughout their entire life cycle according to claim 2, characterized in that, When normalizing the first and second difference features, the range normalization method is used to map the difference and ratio to the [0,1] interval and then re-encode them.

4. The method for managing environmental protection engineering projects throughout their entire life cycle according to claim 2, characterized in that, The specific steps for inputting the first and second difference features into the pre-constructed deviation propagation analysis model to calculate the cumulative impact weight of the design deviation during the construction phase and the attenuation coefficient of the construction deviation during the operation and maintenance phase are as follows: The first difference feature is input as an input variable to the first-level calculation node of the deviation propagation analysis model. The first-level calculation node has a pre-stored set of construction process influence factors. The first difference feature is weighted and aggregated with each factor in the set of construction process influence factors to obtain the cumulative influence weight of the design deviation in the construction stage. The second difference feature is input as an input variable to the second-level calculation node of the deviation propagation analysis model. The second-level calculation node has a pre-stored set of environmental impact factors. The second difference feature is weighted and aggregated with each factor in the set of environmental impact factors to obtain the initial attenuation coefficient of construction deviation in the operation and maintenance stage. The cumulative impact weight is fed back to the second-level calculation node to correct the initial attenuation coefficient, thus obtaining the final attenuation coefficient of the construction deviation during the operation and maintenance phase.

5. The method for managing environmental protection engineering projects throughout their entire life cycle according to claim 4, characterized in that, When feeding the cumulative impact weight back to the second-level calculation node, a multiplicative correction method is used, whereby the cumulative impact weight is used as a multiplier of the attenuation factor to scale the initial attenuation coefficient.

6. The full life cycle environmental engineering project management method according to claim 4, wherein, The specific steps for reverse correction of the monitoring parameters during the operation and maintenance phase based on the cumulative impact weight and attenuation coefficient to generate the corrected operation and maintenance baseline parameters are as follows: Extract the current values ​​of the monitoring parameters during the operation and maintenance phase from the multidimensional project data map, and multiply the current values ​​by the cumulative impact weight to obtain the weighted monitoring values. Dividing the weighted monitoring values ​​by the attenuation coefficient yields the preliminary inverse-calculated equivalent parameters for the construction stage. The equivalent parameters of the construction stage derived from the initial back-inference are fused and calibrated with the measured parameters of the construction stage stored in the multi-dimensional project data map, and the fused and calibrated results are used as the corrected operation and maintenance benchmark parameters.

7. The full life cycle environmental engineering project management method according to claim 6, wherein, The specific steps for comparing the revised operation and maintenance baseline parameters with the design parameters from the project initiation stage to determine whether a project design correction instruction needs to be triggered are as follows: Calculate the absolute and relative deviations between the corrected operation and maintenance baseline parameters and the design parameters from the project initiation stage; The absolute deviation is compared with a preset absolute deviation threshold, and the relative deviation is compared with a preset relative deviation threshold. When the absolute deviation exceeds the absolute deviation threshold and the relative deviation exceeds the relative deviation threshold, a project design correction instruction containing deviation parameter types is generated and pushed to the project design end. When the absolute deviation does not exceed the absolute deviation threshold or the relative deviation does not exceed the relative deviation threshold, no project design correction instruction will be generated, and the monitoring parameters of the current operation and maintenance phase will be marked as acceptable.

8. The full life cycle environmental engineering project management method according to claim 7, wherein, When a project design correction instruction containing deviation parameter types is pushed to the project design end, a comparison curve between the corrected operation and maintenance baseline parameters and the design parameters in the project initiation stage is also pushed.

9. A full life cycle environmental engineering project management system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the full life cycle environmental protection engineering project management method as described in any one of claims 1 to 8.

Images