Civil engineering cost evaluation method and system based on model optimization

Abstract

The invention discloses a civil engineering cost evaluation method and system based on model optimization, and particularly relates to the technical field of civil engineering cost evaluation. Three-dimensional building information and the actual environment of a construction site are combined by establishing a virtual simulation scene based on a BIM model and combining a virtual reality or augmented reality technology, meanwhile, environment variables related to actual construction are dynamically introduced, the variables are analyzed through a machine learning algorithm, risk assessment is conducted on different construction stages, and the risk assessment efficiency is improved. The BIM model is dynamically adjusted and optimized, and based on the optimized BIM model and the risk assessment result, cost assessment is performed for different construction schemes and a report is output, so that the difference between a traditional virtual environment and a real construction condition is effectively made up, the construction cost and risk can be accurately predicted, project hyperbranched and delay are reduced, and the construction efficiency is improved. And more accurate cost evaluation and construction management are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of civil engineering cost assessment, and in particular to a civil engineering cost assessment method and system based on model optimization. Background Art

[0002] In civil engineering projects, cost assessment is one of the important links in project management. Accurate cost assessment can not only help all parties in the project control the budget and reduce cost risks, but also optimize resource allocation and improve the overall benefits of the project. Traditional cost assessment methods mainly rely on manual experience, historical data and quantitative analysis. Although these methods are feasible to a certain extent, their limitations are gradually emerging with the increase in the scale and complexity of the project. For example, traditional methods often find it difficult to effectively handle large amounts of diverse and interrelated engineering data, which affects the accuracy and timeliness of the assessment results. In addition, existing methods find it difficult to fully consider dynamic factors that affect the cost, such as market price fluctuations, changes in the construction environment, etc., which leads to a lack of flexibility and foresight in long-term engineering projects. As a digital modeling tool, BIM (Building Information Modeling) technology can integrate the data of the entire life cycle of a building on the same platform, so that cost assessment can be based not only on data from the design phase, but also on information from the construction and operation and maintenance phases to achieve dynamic assessment of the entire process.

[0003] The prior art has the following deficiencies: BIM models are often combined with virtual reality (VR) or augmented reality (AR) technology to provide pre-construction simulation and cost assessment. However, perfect construction conditions in a virtual environment may differ greatly from actual construction conditions. For example, the model may assume idealized conditions on the construction site (such as precise material supply, worker efficiency, weather, etc.), but these variables are uncontrollable in reality. If project decision makers rely too much on virtual data in the BIM model and ignore the dynamic changes in reality, it may lead to serious deviations between cost assessment and actual costs, project overruns or delays. In addition, perfect simulations in virtual environments may cause participants to underestimate risks in reality, leading to greater engineering problems. Summary of the invention

[0004] The purpose of the present invention is to provide a civil engineering cost evaluation method and system based on model optimization to solve the shortcomings of the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solution: a civil engineering cost evaluation method based on model optimization, comprising the following steps: S1: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; S2: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. S3: For different construction plans, cost assessment is carried out for each plan based on the optimized BIM model and risk assessment results, and the final cost assessment results are output in the form of a report.

[0006] Preferably, in S1, the coordinate systems of the building information in the BIM model and the three-dimensional environmental data of the site are aligned and converted, and the three-dimensional model of the building is embedded in the actual terrain data of the construction site to ensure that the building matches the surrounding natural or man-made environment; the matched BIM model and the construction site environment are imported into the virtual reality platform, the BIM model is superimposed on the real-time picture of the construction site, and the construction simulation tools in the virtual environment are used to simulate each step of the construction process based on the construction schedule in the BIM model.

[0007] Preferably, in S2, in the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced, and the environmental variables include a material supply deviation index and a construction efficiency fluctuation index; Among them, the method for obtaining the material supply deviation index is: within s time periods, compare the expected material supply with the actual material supply, calculate the material supply deviation at each time point, and find the minimum cumulative supply deviation value by optimizing the supply strategy. At t time points, compare the previous status and the current supply situation, dynamically adjust the supply strategy, and select the optimal supply decision based on historical data and current supply capacity to minimize the cumulative supply deviation. The cumulative supply deviations of all time points are added together to obtain the final material supply deviation index, which is used to measure the overall material supply deviation.

[0008] Preferably, in S2, factors affecting construction efficiency need to be determined first, including personnel factors, equipment factors, environmental factors, material supply factors and construction plans; these factors are used as nodes in a Bayesian network to construct a dependency graph, and a conditional probability table is set for each node to represent the probability distribution of different influencing factors under the same conditions. By inputting known conditions of the current time period, the probability distribution of construction efficiency in the current time period is calculated, and then the construction efficiency distribution in the current time period is inferred based on the probability table; the fluctuation value of construction efficiency is obtained by calculating the difference between the actual construction efficiency and the expected construction efficiency in each time period, and the fluctuation values ​​of each time period are accumulated to finally obtain a construction efficiency fluctuation index to measure changes in efficiency during the construction process.

[0009] Preferably, in S2, the material supply deviation index and the construction efficiency fluctuation index are converted into a first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the risk value labels of different construction stages for each group of first eigenvectors as the prediction target, and minimizes the sum of prediction errors for the risk value labels of all different construction stages as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The risk values ​​of different construction stages are determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0010] Preferably, the acquired risk values ​​of different construction stages are compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the risk values ​​of different construction stages are compared with the first standard threshold and the second standard threshold respectively; If the risk value of the construction stage is greater than the second standard threshold, it means that the risk of the construction stage is high. At this time, a first-level warning signal is generated and it is marked as a high-risk construction stage; If the risk value of the construction stage is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the risk of the construction stage is average. At this time, a secondary warning signal is generated and it is marked as a medium-risk construction stage; If the risk value of the construction phase is less than the first standard threshold, it means that the risk of the construction phase is low. At this time, a third-level warning signal is generated and it is marked as a low-risk construction phase.

[0011] Preferably, historical data related to different construction schemes are collected, data cleaning is performed, missing values ​​and outliers are processed, and the data are divided into training sets and test sets, with 70%-80% used for training and 20%-30% used for testing; the random forest model consists of multiple decision trees, and the model constructs each tree by randomly selecting samples and features from the training set. Each decision tree learns from random samples and independently generates prediction results. Finally, the random forest determines the final prediction value by averaging or voting the prediction values ​​of multiple decision trees. The features of the training set are input into the model, and the random forest finds the optimal split point by continuously splitting samples to generate rules for predicting the cost; after the model training is completed, it is evaluated using the test set, and the trained random forest model is applied to the cost prediction of the new construction scheme; the features of the new scheme are input into the model, and the model will generate prediction values ​​based on the previously learned rules using each decision tree, and then output the total cost prediction by averaging these prediction values; the prediction process of the model is to use all the trained decision trees, and each tree independently generates a prediction value. Finally, the random forest gives the total cost prediction of the new construction scheme by the average of these prediction values.

[0012] The present invention also provides a civil engineering cost evaluation system based on model optimization, including a virtual simulation scene generation module, an engineering risk evaluation module and an evaluation report output module; Virtual simulation scene generation module: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; Engineering risk assessment module: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. Evaluation report output module: For different construction plans, cost evaluation is carried out for each plan according to the optimized BIM model and risk assessment results, and the final cost evaluation results are output in the form of a report.

[0013] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention combines the BIM model with virtual reality or augmented reality technology to generate a virtual simulation scene before construction, and introduces dynamic environmental variables in actual construction (such as material supply deviation index and construction efficiency fluctuation index). Based on these environmental variables, risk assessment is performed on different construction stages through machine learning algorithms, and the BIM model is dynamically adjusted and optimized based on the assessment results. This process effectively avoids the difference between the virtual simulation scene and the actual construction conditions, ensures that the cost assessment is closer to reality, and reduces the risk of project overruns and delays.

[0014] 2. The present invention accurately predicts the cost of different construction plans by introducing machine learning models such as polynomial regression and random forest. By classifying and warning the risk values ​​of the construction stage, it can identify high, medium and low risk construction stages and provide timely risk warnings for project managers, effectively solving the error problem caused by over-reliance on virtual environments in the prior art, and can more accurately evaluate construction costs, ensuring that the overall progress and budget of the project are within a reasonable range. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 The figure is a flow chart of the method of the present invention.

[0017] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1, please refer to Figure 1 and Figure 2 As shown, the civil engineering cost evaluation method based on model optimization described in this embodiment includes the following steps: S1: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; S2: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. S3: For different construction plans, cost assessment is carried out for each plan based on the optimized BIM model and risk assessment results, and the final cost assessment results are output in the form of a report.

[0020] Among them, in S1, according to the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction, specifically: Based on the design drawings, building specifications, material information and construction plans of civil engineering projects, a BIM model is established. The model should contain detailed data from multiple professional fields such as the building's three-dimensional geometric information, structural system, material properties, mechanical and electrical equipment layout, construction progress, etc.

[0021] Ensure that all components and design details in the BIM model are consistent with the actual requirements of the project, and optimize the details as needed to ensure that it can be seamlessly integrated into the virtual environment. The file format of the model should be compatible with the virtual reality or augmented reality platform, and commonly used file formats include IFC, Revit, Navisworks, etc.

[0022] Utilize technologies such as drones, laser scanning (LiDAR) or geographic information systems (GIS) to collect detailed geographic information and three-dimensional terrain data of the construction site, including ground elevation differences, the location and height of surrounding buildings, and natural environment characteristics. Obtain dynamic data of the construction site, such as real-time weather, lighting conditions, surrounding traffic conditions, and surface conditions. These data can be collected through sensor networks, meteorological service platforms, and IoT devices to reflect the actual construction environment during the simulation process.

[0023] Align and convert the coordinate system of the building information in the BIM model with the 3D environment data on site. By unifying the spatial coordinate systems of the two, ensure that the location of the building in the virtual scene is consistent with the physical location in reality.

[0024] Embed the 3D model of the building into the actual terrain data of the construction site to ensure that the building matches the surrounding natural or man-made environment. For example, the foundation of the building should be highly consistent with the actual terrain of the site, and the spatial relationship around the building should be consistent with the site conditions.

[0025] Consider the shading effects and lighting conditions of buildings in the site environment. Use lighting simulation tools to analyze the changes in shadows and the impact of sunlight on buildings during construction or after completion to ensure that the model reflects real-world lighting conditions.

[0026] The matched BIM model and the construction site environment are imported into the virtual reality platform, such as Unity, Unreal Engine and other 3D engines, to generate interactive virtual scenes. In virtual reality devices (such as VR helmets, CAVE systems), users can immersively view the combination of the building's 3D model and the construction site, and conduct advance construction planning and cost assessment.

[0027] The BIM model is superimposed on the real-time picture of the construction site through augmented reality devices (such as Microsoft HoloLens, iPad and other AR devices). On-site managers can use AR devices to view the dynamic matching between the building model and the real environment in real time, and virtually "preview" the spatial layout and construction process of the building on site.

[0028] Using construction simulation tools in a virtual environment, based on the construction schedule in the BIM model, simulate each step in the construction process, including earth excavation, structural construction, material transportation and installation, etc. Combined with real-time data from the construction site, verify the feasibility of the construction schedule and analyze possible obstacles or delays.

[0029] Through real-time collision detection in the virtual scene, the feasibility of the building design and construction plan can be verified in the on-site environment. Any potential conflicts with surrounding facilities, terrain or traffic on site can be identified and adjusted in advance in the simulation.

[0030] In the virtual simulation scene, the construction team, designers and cost estimators can collaborate to review the rationality of the integration of the BIM model with the site and put forward suggestions for improvement. If it is found that the building design or construction plan in the virtual environment does not match the actual conditions, the team can adjust the BIM model.

[0031] According to the feedback from the construction team, the BIM model and construction plan are optimized. For example, the material supply time is adjusted, the equipment layout is optimized, the construction sequence is modified, etc. After the adjustment is completed, the virtual simulation scene is regenerated and the improvement effect is verified.

[0032] After completing the matching, adjustment and optimization of the BIM model and the construction environment, the final pre-construction virtual simulation scene is generated. This scene should reflect the actual situation of the construction site and dynamically update the real-time data in the model.

[0033] The final virtual simulation scene is used to train the construction team and instruct them on how to implement the plan on the construction site. Through virtual reality or augmented reality equipment, construction personnel can familiarize themselves with the site environment and building layout in advance to ensure accuracy and efficiency during the construction process.

[0034] In the final virtual simulation scenario, cost assessment and risk analysis are performed based on the dynamic data of the BIM model and the simulation of the on-site environment. Any potential delays, cost overruns or resource allocation issues can be identified and adjusted in advance in the virtual scenario.

[0035] During the construction process, IoT devices and on-site sensors are used to continuously collect real-time data from the construction site and dynamically update the BIM model in the virtual simulation scene. This includes the synchronization of data such as construction progress, personnel allocation, weather conditions, and equipment status.

[0036] As construction progresses, the virtual scene should continue to be consistent with the actual environment, and the construction progress should be monitored through continuous virtual reality simulation to ensure that the entire project is implemented within the predetermined cost and time frame.

[0037] In this application, the three-dimensional building information in the BIM model can be combined with the actual environment of the construction site to generate a virtual construction simulation scene that includes dynamic data and virtual reality / augmented reality integration. This scene not only helps the early planning and cost evaluation of the project, but also provides real-time feedback and optimization during the construction process, realizing dynamic monitoring and management of the construction site.

[0038] S2: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results.

[0039] In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced, including the material supply deviation index and the construction efficiency fluctuation index; The material supply deviation index is obtained by marking the expected material supply at time t as , marking the actual material supply as , calculate the supply deviation , the expression is: ; Accumulate to the current minimum supply deviation through the optimal supply strategy , the expression is: ; Calculate decision variables , represents the supply strategy decision at time t, which determines the actual amount of material supplied in the tth time period , the formula is: ; The current state The status at the previous time point For comparison, construct the state transfer equation: ; In the dynamic programming calculation, the boundary condition is the initial supply state. In the first time period t=1, the accumulated supply deviation is: ; For each time period t (from 2 to s), the sum of supply deviations is iteratively updated according to the state transition equation, expressed as: ; In each time period, select the optimal supply strategy based on historical supply conditions and current supply capabilities , so that At the end of all time periods s, the cumulative sum of supply deviations is the final material supply deviation index, which is expressed as: ; Where MS is the material supply deviation index.

[0040] The larger the Material Supply Deviation Index (MSDI), the greater the difference between the actual amount of materials supplied and the expected plan. This usually indicates that the uncertainty and instability of material supply during the construction process has increased, resulting in a significant increase in potential construction risks. A large supply deviation index may be due to a variety of factors, such as suppliers' inability to deliver on time, delays in transportation, insufficient material inventory, or even material shortages caused by market fluctuations. These situations can lead to interruptions or delays in the supply of materials at the construction site, which directly affects the progress of the project.

[0041] When the supply deviation index is large, the project team needs to frequently respond to emergencies, such as temporarily finding new suppliers or alternative materials, which increases the complexity of construction management. In addition, due to the uncertainty of material supply at the construction site, construction delays may occur, resulting in idle workers or inability of construction equipment to operate normally. This not only adds additional operating costs, but may also affect the overall completion time of the project. More seriously, frequent material supply problems may affect the quality of the building, especially when key materials cannot be obtained in time, which may force the construction party to use inferior quality alternative materials, thereby increasing future repair and maintenance costs. Therefore, the larger the material supply deviation index, the higher the risk of the construction phase, and the project progress faces more uncertainties and uncontrollable factors.

[0042] On the contrary, when the material supply deviation index is small, it means that the actual supplied materials are basically consistent with the planned demand, and the supply chain is relatively stable and reliable. This shows that during the construction process, the supply of materials has a high degree of predictability and certainty, and the materials required by the construction site can be supplied on time and in quantity, reducing the risks caused by material shortages or delays in the construction progress. In this case, the project can proceed smoothly according to the predetermined plan, and the workers and equipment can continue to operate efficiently, without stopping work or reducing efficiency due to insufficient materials.

[0043] A low material supply deviation index usually indicates that the supplier has strong delivery capabilities, the logistics and transportation process is smooth, and the project manager can effectively manage inventory and arrange emergency plans. With the stability of material supply, the risk of construction progress is reduced, and the construction party can better control the cost and time of the project and ensure that the quality is up to standard. This low-deviation material supply situation not only reduces the management pressure of the project, but also enhances the construction team's confidence in the smooth completion of the project, thereby reducing the potential quality and time risks during the construction process. Therefore, the smaller the material supply deviation index, the lower the risk of the construction phase, and the more robust and controllable the project as a whole.

[0044] Among them, the method for obtaining the construction efficiency fluctuation index is: determine the factors that affect the construction efficiency. Construction efficiency is the result of the joint action of multiple factors, including: personnel factors: the number of workers, experience, fatigue status, attendance rate, etc. Equipment factors: equipment availability, failure rate, maintenance frequency, etc. Environmental factors: weather conditions (such as temperature, rainfall, wind speed, etc.), construction site conditions. Material supply: timely supply of materials, supply chain delays, etc. Construction plan: task scheduling, process arrangement, etc., take these factors as nodes in the Bayesian network, each node represents a random variable, construct the structure of the Bayesian network, and represent the dependencies between the influencing factors with a directed acyclic graph (DAG). For each node (influencing factor) and their parent nodes, define a conditional probability table (CPT). The conditional probability table represents the probability distribution of the value a node takes given its parent node. For example, construction efficiency Depends on multiple parent nodes, such as personnel factors and equipment factors Environmental factors , Material Supply Etc., then the conditional probability table of construction efficiency is: ; h is time; In the Bayesian network, by setting known inputs (such as current weather conditions, material supply, staffing, etc.), the probability distribution of construction efficiency at the current time point is calculated. The known input variables of the current time period (such as weather data, number of workers, equipment status) are substituted into the Bayesian network. According to the conditional probability table, the probability distribution of construction efficiency at the current time period is inferred. Probability distribution of construction efficiency , the expression is: ;in, are different levels of construction efficiency (e.g., low efficiency, medium efficiency, high efficiency). Based on the probability distribution of construction efficiency, calculate the construction efficiency fluctuation value for the current time period The fluctuation value can be measured by the difference between the expected value of construction efficiency and the actual value of current construction efficiency, expressed as: ; is the actual construction efficiency in the hth time period, The expected construction efficiency inferred from the Bayesian network is usually the expected value of the probability distribution. ; Accumulate the fluctuation values ​​of multiple time periods to obtain the construction efficiency fluctuation index, which reflects the degree of efficiency fluctuation in the entire construction process. ; T is the total number of time periods, is the construction efficiency fluctuation index.

[0045] The larger the construction efficiency fluctuation index, the greater the fluctuation of construction efficiency between different time periods, and the poor stability of the construction stage. Such large fluctuations usually mean that various uncertain factors at the construction site have a significant impact on efficiency, such as unstable worker attendance, frequent equipment failures, untimely material supply or weather changes. These fluctuations lead to large deviations between the planning and actual conditions during the construction process, which may lead to problems such as construction delays and cost overruns. Since the construction efficiency cannot be kept stable, it is difficult for managers to predict the next progress and resource requirements in advance, which will increase the difficulty and uncontrollability of management during the construction process. Large efficiency fluctuations may also lead to inefficient use of workers and equipment, further exacerbating construction risks and even affecting the quality of the project. Therefore, the larger the construction efficiency fluctuation index, the higher the risk of the construction stage, and the more uncertain the completion time and cost control of the project face.

[0046] On the contrary, when the construction efficiency fluctuation index is small, it indicates that the efficiency fluctuation in the construction process is small, and the overall efficiency remains at a relatively stable level. Small fluctuations in construction efficiency usually mean that the construction plan is well executed, key factors such as worker attendance, equipment operation, and material supply are relatively stable, and environmental factors such as weather conditions have little impact on the construction progress. In this case, project managers can more accurately predict the construction progress and resource requirements, reduce adjustments and delays caused by emergencies, and thus improve the controllability and management efficiency of the project. Stable construction efficiency also helps ensure that the project proceeds as planned and reduces the risk of cost overruns and construction delays. Therefore, the smaller the construction efficiency fluctuation index, the lower the risk of the construction phase, and the smoother and easier to manage the construction process.

[0047] The material supply deviation index and the construction efficiency fluctuation index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the risk value labels of different construction stages with each group of first eigenvectors as the prediction target, and minimizes the sum of prediction errors of risk value labels of all different construction stages as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The risk values ​​of different construction stages are determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0048] The method for obtaining the risk values ​​of different construction stages is as follows: from the first eigenvector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, MS is the material supply deviation index, is the construction efficiency fluctuation index, is the risk value at different construction stages.

[0049] The obtained risk values ​​of different construction stages are compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the risk values ​​of different construction stages are compared with the first standard threshold and the second standard threshold respectively; If the risk value of the construction stage is greater than the second standard threshold, it means that the risk of the construction stage is high. At this time, a first-level warning signal is generated and it is marked as a high-risk construction stage; If the risk value of the construction stage is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the risk of the construction stage is average. At this time, a secondary warning signal is generated and it is marked as a medium-risk construction stage; If the risk value of the construction phase is less than the first standard threshold, it means that the risk of the construction phase is low. At this time, a third-level warning signal is generated and it is marked as a low-risk construction phase.

[0050] It should be noted here that the importance of the first-level warning signal is greater than that of the second-level warning signal, and the importance of the second-level warning signal is greater than that of the third-level warning signal. Relevant personnel can take corresponding adjustment measures according to different warning signal levels.

[0051] When the risk value of the construction phase is greater than the second standard threshold, it indicates that the risk of this phase is high and a first-level warning signal is generated. At this time, urgent adjustment and optimization measures need to be taken to reduce major risks.

[0052] a) Optimize material supply chain management Strengthen coordination with suppliers: In the case of excessively high material supply deviation index, it is necessary to optimize the supply chain and contact suppliers of key materials to ensure that materials are supplied on time and reduce the possibility of delays. The inventory of key materials can be increased, or alternative suppliers can be sought to disperse supply risks.

[0053] Dynamically adjust inventory management: Through the real-time supply data and construction plan integrated in the BIM model, adjust the inventory quantity and supply time of materials to ensure that there is no shortage of important materials during construction.

[0054] Real-time monitoring of supply chain nodes: Use the Internet of Things (IoT) and sensors to monitor the status of each link in the supply chain, grasp the actual progress of material supply in real time, detect problems in advance and take countermeasures.

[0055] b) Optimize personnel and equipment scheduling Increase staffing: If the construction efficiency fluctuation index is too high, the construction efficiency can be improved by increasing the number of workers or introducing more experienced personnel. Based on real-time efficiency data, reallocate personnel to ensure construction progress.

[0056] Equipment maintenance and optimization: Check the operating status of key equipment, arrange necessary maintenance or replacement, and reduce the impact of equipment failure on construction efficiency. Use the equipment usage history data in the BIM model to analyze the equipment failure rate and arrange maintenance plans in advance.

[0057] c) Dynamically adjust construction plans Adjust the construction sequence: For high-risk construction stages, dynamically adjust the construction plan and process sequence, give priority to high-risk parts, and advance or postpone some work to optimize resource utilization.

[0058] Refined progress control: In high-risk stages, introduce more fine-grained progress control methods, monitor construction progress in real time, and make dynamic adjustments based on the status of materials and equipment.

[0059] When the risk value of the construction stage is between the first standard threshold and the second standard threshold, a second-level warning signal is generated, indicating that there is a certain risk in the construction stage, but it has not yet reached the high risk level. At this time, preventive adjustments are required to ensure that the construction risk is effectively controlled.

[0060] a) Preventive measures Monitor material supply time: Although the material supply deviation index has not reached an extreme level, in the medium-risk stage, the early warning mechanism for material supply should be strengthened, and communication with suppliers should be carried out in advance to ensure that material supply will not be delayed. The dynamic supply information of materials in the BIM model can be used to set a reasonable lead time.

[0061] b) Personnel and equipment optimization Adjust staffing strategies: In the medium-risk stage, the scheduling of workers and equipment can be optimized appropriately. Through real-time data in the BIM model, identify which processes may require more resource support, and deploy manpower and equipment in advance to ensure smooth construction.

[0062] Backup plan: Develop backup plans in case of equipment failure, such as preparing backup equipment or scheduling equipment maintenance in advance, to reduce the impact of equipment failure on efficiency.

[0063] c) Process adjustment and acceleration Accelerate key processes: While ensuring construction quality, speed up the progress of key processes to reduce pressure in subsequent stages. Use the construction progress simulation function in the BIM model to evaluate the impact of accelerated construction on the overall project and make adjustments based on the simulation results.

[0064] When the risk value of the construction phase is less than the first standard threshold, a third-level warning signal is generated, indicating that the construction risk at this stage is low and the focus can be on efficiency improvement and cost optimization.

[0065] a) Efficiency improvement and cost control Optimize resource allocation: In the low-risk stage, the use of some personnel and equipment can be reduced to avoid resource waste. Through the resource usage information in the BIM model, optimize resource allocation and allocate more resources to the construction stage with higher risks.

[0066] Streamlining the process: Analyze the redundant steps in the construction process, simplify or optimize them, and improve construction efficiency. In the BIM model, the construction process is optimized through virtual simulation and applied to actual construction.

[0067] b) Continuously monitor and prevent risks Real-time data analysis: Even at a low-risk stage, the real-time monitoring function in the BIM model should be used continuously to analyze construction efficiency and material supply to ensure that timely response measures can be taken when potential problems arise to prevent risks from escalating.

[0068] Stay connected to the early warning system of external factors: monitor the potential impact of changes in the external environment (such as weather, regulatory changes, etc.) on low-risk construction stages, and dynamically update the construction plan to respond to possible external risks.

[0069] 4. Dynamic adjustment of BIM models During the entire construction process, combined with early warning signals at different stages, the BIM model needs to dynamically update data and make adjustments to reflect the latest construction status.

[0070] a) Real-time data update Real-time update of material supply and equipment status: At each construction stage, the material inventory, supply time and equipment operation status in the BIM model are updated in real time based on the on-site IoT sensor data and supply chain information to ensure that the construction information is always up to date.

[0071] b) Dynamic adjustment of construction progress Adjust the construction plan based on risk: Adjust the construction plan based on the risk value of different construction stages, focusing on progress and resource allocation in high-risk stages, and focusing on cost control and efficiency improvement in low-risk stages. Use the construction simulation function in the BIM model to verify the impact of the adjusted plan on the overall progress.

[0072] c) Decision support and risk management Use BIM models to support decision-making: By integrating predicted data on construction risks, BIM models can provide decision-making support for project managers. Dynamically adjust the construction plan, resource allocation, and risk management strategies in the model to ensure that risks at each construction stage are effectively controlled.

[0073] In this application, by analyzing the material supply deviation index and the construction efficiency fluctuation index, combined with the risk values ​​of different construction stages predicted by the machine learning model, the BIM model can be dynamically adjusted and optimized for different risk levels. For the high-risk stage, the focus is on optimizing the supply chain, improving resource allocation efficiency and adjusting the construction plan; for the medium-risk stage, preventive measures are taken and scheduling is optimized; for the low-risk stage, the focus is on efficiency improvement and cost control. Through continuous monitoring and adjustment, the BIM model can ensure the smooth progress of the construction project, reduce risks and optimize the overall construction process.

[0074] S3: For different construction plans, cost assessment is carried out for each plan based on the optimized BIM model and risk assessment results, and the final cost assessment results are output in the form of a report.

[0075] First, it is necessary to collect historical data related to different construction plans. The main data includes: Material cost: the unit price, usage and type of each material. For example, the price and usage of cement, steel bars, bricks and stones.

[0076] Labor costs: hourly wages and working hours for different types of work. Including labor costs of ordinary workers, skilled workers, etc.

[0077] Equipment costs: The unit price of renting or purchasing equipment, the length of time it is used, and the cost of maintenance. For example, the rental cost of a crane or excavator and how long it is used.

[0078] Risk assessment data: risk-related indicators, such as material supply deviation index, construction efficiency fluctuation index, etc. These risk assessment results help the model better understand the uncertainties in construction projects.

[0079] Other characteristics: such as geographical location, weather conditions, project size, etc., will also have an impact on construction costs.

[0080] Data cleaning includes dealing with missing values ​​and outliers. If there are missing values ​​in the material unit price or labor time data, you may need to fill them with the mean or median, or delete the records that cannot be corrected. In addition, categorical data (such as worker category, equipment type, etc.) need to be converted into numerical form for input into the model. For example, the worker category can be represented by one-hot encoding.

[0081] The input variables of the model, such as material cost, labor cost, equipment cost, risk assessment results, etc. The target variable to be predicted by the model, that is, the total cost of each construction plan.

[0082] After data preparation, the dataset is divided into two parts: Training set: used to train the model, usually accounting for 70%-80% of the total data.

[0083] Test set: used to verify model performance, usually accounting for 20%-30% of the total data.

[0084] The training set is used to allow the model to learn the cost structure and patterns of different construction options, and the test set is used to evaluate the model's generalization ability, that is, its performance when processing unseen data.

[0085] The random forest model consists of multiple decision trees, each of which is built independently. The random forest builds each tree by randomly selecting samples from the training set and randomly selecting features each time. The main parameters in the model include: n_estimators: The number of decision trees. The more trees there are, the better the prediction stability and accuracy of the model.

[0086] max_features: The maximum number of features used by each tree when splitting. By limiting the number of features per tree, you can reduce overfitting of the model.

[0087] max_depth: The maximum depth of each tree. Limiting the depth of the tree can prevent the model from overfitting.

[0088] Each decision tree learns from a random sample of the training set and generates predictions independently. The final prediction is determined by the average or voting results of the predictions of all trees. This approach improves the performance and robustness of the overall model by combining multiple weak learners (i.e., multiple decision trees).

[0089] The features (material, labor, equipment, etc. costs) and labels (total cost of the construction plan) in the training set are input into the model for training. Random forest will generate multiple decision trees based on different samples and features. Each tree repeatedly splits the data to find the optimal split point and gradually generate rules for predicting the total cost.

[0090] By using multiple trees, the model can reduce the bias and overfitting problems that a single tree may bring. The more trees there are, the better the model fits the training data, and the better it can cope with the uncertainty in the test data.

[0091] After training is completed, the model is evaluated using the test set. The evaluation indicators mainly include: Mean Squared Error (MSE): It is used to measure the average square difference between the model prediction value and the true value. The larger the difference, the more inaccurate the model prediction is.

[0092] Root Mean Squared Error (RMSE): The square root of the mean square error, which more intuitively represents the average prediction error of the model.

[0093] Mean Absolute Error (MAE): Indicates the average absolute difference between the predicted value and the true value. Compared with MSE, MAE is more intuitive and is not affected by extreme errors.

[0094] Model evaluation can effectively measure the performance of the model through these indicators and help determine the accuracy of the model. If the test results have large errors, it may be necessary to adjust the model parameters (such as the number of trees, maximum depth, etc.) or perform further data processing.

[0095] The trained random forest model can be used to predict the cost of a new construction plan. For a new construction plan, input its feature information (such as material cost, labor cost, equipment requirements, etc.), and the model will output the total cost forecast of the plan based on the previously learned rules.

[0096] The model prediction process is to pass the data of the new construction plan through all the trained decision trees, each tree generates a prediction value, and finally the random forest generates the final total cost prediction by the average of these prediction values.

[0097] Based on the prediction results of the model, a cost assessment report for each construction plan can be generated. The report includes: Total cost of the construction plan: the final total cost of the construction plan predicted by the model.

[0098] Material cost breakdown: includes the quantity, unit price and total cost of each material.

[0099] Labor cost breakdown: including job classification, working hours and total labor cost.

[0100] Equipment cost breakdown: equipment usage time, rental fees, maintenance costs, etc.

[0101] Risk-adjusted costs: Based on the risk assessment results learned by the model, it reflects the increase or decrease in costs caused by deviations in material supply, fluctuations in construction efficiency, etc.

[0102] Other costs: such as management fees, unforeseen expenses, etc.

[0103] In this application, the steps of cost assessment through random forest model include data preparation, model construction, model training, evaluation and prediction. The random forest model generates accurate and robust prediction results through the combination of multiple decision trees, which is particularly suitable for complex multi-factor cost assessment problems. The model can process multidimensional data such as materials, labor, equipment and risk assessment, and finally output the total cost of the construction plan and generate a complete cost assessment report. This method can help project managers better control budgets and identify potential risks, and improve the cost prediction ability of construction projects.

[0104] In this embodiment, according to the design drawings of the civil engineering project, the full life cycle information of the building is first established through the BIM model, and the three-dimensional building information in the BIM model is combined with the actual environment of the construction site by using virtual reality or augmented reality technology to generate a virtual simulation scene before construction. In the simulation process of this virtual scene, environmental variables related to the actual construction are dynamically introduced, and these variables are analyzed by machine learning algorithms, and then risk assessment is performed for different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. Finally, for different construction plans, a detailed cost assessment is performed based on the optimized BIM model and risk assessment results, and a cost assessment report for each construction plan is output to assist decision-making.

[0105] Embodiment 2, the civil engineering cost assessment system based on model optimization described in this embodiment includes a virtual simulation scene generation module, an engineering risk assessment module and an assessment report output module; Virtual simulation scene generation module: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; Engineering risk assessment module: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. Evaluation report output module: For different construction plans, cost evaluation is carried out for each plan according to the optimized BIM model and risk assessment results, and the final cost evaluation results are output in the form of a report.

[0106] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0107] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0108] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0109] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A civil engineering cost assessment method based on model optimization, characterized by: The steps include: S1: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; S2: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. S3: For different construction plans, cost assessment is carried out for each plan based on the optimized BIM model and risk assessment results, and the final cost assessment results are output in the form of a report.

2. The civil engineering cost assessment method based on model optimization according to claim 1 is characterized in that: In S1, the coordinate systems of the building information in the BIM model and the three-dimensional environmental data on site are aligned and converted, and the three-dimensional model of the building is embedded in the actual terrain data of the construction site to ensure that the building matches the surrounding natural or man-made environment; the matched BIM model and the construction site environment are imported into the virtual reality platform, the BIM model is superimposed on the real-time picture of the construction site, and the construction simulation tools in the virtual environment are used to simulate each step of the construction process based on the construction schedule in the BIM model.

3. The civil engineering cost assessment method based on model optimization according to claim 2 is characterized in that: In S2, during the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced, and the environmental variables include the material supply deviation index and the construction efficiency fluctuation index; Among them, the method for obtaining the material supply deviation index is: within s time periods, compare the expected material supply with the actual material supply, calculate the material supply deviation at each time point, and find the minimum cumulative supply deviation value by optimizing the supply strategy. At t time points, compare the previous status and the current supply situation, dynamically adjust the supply strategy, and select the optimal supply decision based on historical data and current supply capacity to minimize the cumulative supply deviation. The cumulative supply deviations of all time points are added together to obtain the final material supply deviation index, which is used to measure the overall material supply deviation.

4. The civil engineering cost assessment method based on model optimization according to claim 3 is characterized by: In S2, we first need to determine the factors that affect construction efficiency, including personnel factors, equipment factors, environmental factors, material supply factors and construction plans; use them as nodes in the Bayesian network, build a dependency graph, and set a conditional probability table for each node to represent the probability distribution of different influencing factors under the same conditions. By inputting the known conditions of the current time period, we calculate the probability distribution of construction efficiency in the current time period, and then infer the construction efficiency distribution of the current time period based on the probability table; by calculating the difference between the actual construction efficiency and the expected construction efficiency in each time period, we get the fluctuation value of construction efficiency, and accumulate the fluctuation values ​​of each time period to finally get the construction efficiency fluctuation index, which is used to measure the changes in efficiency during the construction process.

5. The civil engineering cost assessment method based on model optimization according to claim 4 is characterized in that: In S2, the material supply deviation index and the construction efficiency fluctuation index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the risk value labels of different construction stages with each group of first eigenvectors as the prediction target, and minimizes the sum of prediction errors of risk value labels of all different construction stages as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The risk values ​​of different construction stages are determined according to the model output results, wherein the machine learning model is a polynomial regression model.

6. The civil engineering cost assessment method based on model optimization according to claim 5 is characterized by: The obtained risk values ​​of different construction stages are compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the risk values ​​of different construction stages are compared with the first standard threshold and the second standard threshold respectively; If the risk value of the construction stage is greater than the second standard threshold, it means that the risk of the construction stage is high. At this time, a first-level warning signal is generated and it is marked as a high-risk construction stage; If the risk value of the construction stage is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the risk of the construction stage is average. At this time, a secondary warning signal is generated and it is marked as a medium-risk construction stage; If the risk value of the construction phase is less than the first standard threshold, it means that the risk of the construction phase is low. At this time, a third-level warning signal is generated and it is marked as a low-risk construction phase.

7. The civil engineering cost assessment method based on model optimization according to claim 1 is characterized in that: Collect historical data related to different construction plans, clean the data, process missing values ​​and outliers, and divide them into training sets and test sets, with 70%-80% used for training and 20%-30% used for testing; the random forest model consists of multiple decision trees. The model constructs each tree by randomly selecting samples and features from the training set. Each decision tree learns from random samples and independently generates prediction results. Finally, the random forest determines the final prediction value by averaging or voting the prediction values ​​of multiple decision trees. The features of the training set are input into the model. The random forest finds the optimal split point by continuously splitting samples and generates rules for predicting the cost. After the model training is completed, it is evaluated using the test set, and the trained random forest model is applied to the cost prediction of the new construction plan. The features of the new plan are input into the model. The model will generate prediction values ​​based on the previously learned rules using each decision tree, and then output the total cost prediction by averaging these prediction values. The prediction process of the model is to use all the trained decision trees, each tree independently generates a prediction value, and finally the random forest gives the total cost prediction of the new construction plan by the average of these prediction values.

8. A civil engineering cost evaluation system based on model optimization, used to implement the civil engineering cost evaluation method based on model optimization as claimed in any one of claims 1 to 7, characterized in that: It includes virtual simulation scenario generation module, engineering risk assessment module and assessment report output module; Virtual simulation scene generation module: Based on the design drawings of the civil engineering project, a BIM model containing the information of the entire life cycle of the building is established. Through virtual reality or augmented reality technology, the three-dimensional building information in the BIM model is combined with the actual environment of the construction site to generate a virtual simulation scene before construction; Engineering risk assessment module: In the actual simulation process of the virtual simulation scene, environmental variables related to the actual construction are dynamically introduced. Based on the virtual simulation environment, the environmental variables are analyzed through machine learning algorithms. According to the analysis results, risk assessment is performed on different construction stages, and the BIM model is dynamically adjusted and optimized based on the assessment results. Evaluation report output module: For different construction plans, cost evaluation is carried out for each plan according to the optimized BIM model and risk assessment results, and the final cost evaluation results are output in the form of a report.

Images