Artificial intelligence-based prefabricated underground channel field installation construction method

Abstract

The application relates to the field of underground channel construction, and specifically discloses a fabricated underground channel field installation construction method based on artificial intelligence, which comprises the following steps: S1, constructing an initial construction data set; S2, automatically identifying key performance indicators and potential efficiency bottlenecks in construction; S3, generating an optimal construction process and resource allocation scheme by using an improved genetic algorithm; S4, implementing operation instructions and scheduling relevant construction resources by using an artificial intelligence system; S5, constructing a dynamic construction data set; S6, based on the dynamic construction data set collected in real time in step S5, adjusting the construction plan and operation instructions in real time by using the artificial intelligence system; S7, in the design stage of prefabricated components, the artificial intelligence system is applied to predict and simulate the performance of the components; and S8, in the field assembly stage, the prefabricated components are installed under the guidance of the artificial intelligence according to the design scheme in step S7, so that the correct docking and fixing of the components are ensured. The construction time and resource waste are effectively reduced, and the construction efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of culvert construction technology, and in particular to an artificial intelligence-based prefabricated culvert on-site installation method. Background Technology

[0002] In modern infrastructure construction, prefabricated culverts, as a new type of drainage system structure, are widely used due to their advantages such as rapid construction and environmental friendliness. However, in actual construction, the installation of prefabricated culverts faces many challenges, mainly including the complexity of the construction process, the efficiency of resource allocation, the necessity of real-time monitoring and dynamic adjustment, and the precision of prefabricated component design and on-site assembly. Existing technologies often have limitations in addressing these issues, leading to difficulties in achieving the expected construction efficiency and quality.

[0003] Firstly, regarding construction process optimization, traditional construction methods typically rely on experience and manual operation. This approach is inefficient in complex construction environments and makes it difficult to adjust and optimize the construction process in real time. For example, in the construction of prefabricated culverts, issues such as how to effectively arrange various construction tasks, how to rationally allocate construction resources, and how to dynamically adjust the construction plan are all difficult to solve efficiently using traditional methods. Existing technologies lack data-driven intelligent decision support systems, making it impossible to fully utilize real-time data from the construction site for construction process optimization.

[0004] Secondly, in terms of resource allocation and management, existing technologies often struggle to achieve optimal resource configuration. Traditional resource management relies primarily on manual statistics and scheduling, lacking systematic and intelligent management methods. This approach is not only labor-intensive and error-prone, but also fails to enable real-time monitoring and dynamic adjustment of resource usage, leading to resource waste and low construction efficiency. In the construction of prefabricated culverts, how to efficiently allocate manpower, machinery, and materials has become a bottleneck that existing technologies have yet to overcome.

[0005] Furthermore, in terms of real-time monitoring and dynamic adjustment, existing technologies typically rely on manual monitoring and adjustments, lacking systematic and intelligent monitoring methods. The construction environment for prefabricated culverts is complex and variable, often encountering various unforeseen circumstances such as changes in environmental conditions and abnormal material conditions. If these problems are not detected and addressed promptly, they will severely impact construction quality and progress. Traditional monitoring methods often struggle to provide comprehensive real-time monitoring of the construction site, let alone enable timely and effective dynamic adjustments when problems are detected.

[0006] Therefore, how to provide an artificial intelligence-based prefabricated culvert on-site installation method is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] To address the aforementioned issues, this invention provides an artificial intelligence-based method for on-site installation of prefabricated culverts.

[0008] The present invention provides an artificial intelligence-based prefabricated culvert on-site installation construction method, which includes the following steps;

[0009] S1. Collect initial construction data

[0010] Collect location, elevation, water drop, and linear data after the culvert is assembled; calculate the safe hoisting radius and hoisting position based on the hoisting weight; and construct the initial construction dataset.

[0011] S2. Analyze initial construction data

[0012] By applying an improved aggregated decision tree to analyze the initial construction dataset, key performance indicators and potential efficiency bottlenecks in the construction process are automatically identified.

[0013] S3. Generate the optimal construction process and resource allocation plan.

[0014] Based on the analysis results of step S2, an improved genetic algorithm is used to generate the optimal construction process and resource allocation scheme, where resources include manpower, machinery and equipment and materials.

[0015] S4. Convert and implement operation instructions

[0016] The construction process and resource allocation plan generated in step S3 are transformed into specific operation instructions, and the operation instructions are implemented through the artificial intelligence system to schedule relevant construction resources.

[0017] S5. Real-time monitoring of the construction site

[0018] During construction, a variety of sensors are used to monitor the environmental conditions, material status and worker behavior at the construction site in real time, and to build a dynamic construction dataset. The sensors include position sensors, elevation sensors, water drop sensors, sensors for monitoring the linearity of the culvert after assembly, and hoisting safety sensors.

[0019] S6. Dynamically adjust the construction plan.

[0020] Based on the dynamic construction dataset collected in real time in step S5, the construction plan and operation instructions are adjusted in real time through the artificial intelligence system to respond to any deviations or potential problems detected in real time.

[0021] S7, Optimize prefabricated component design

[0022] During the design phase of prefabricated components, artificial intelligence systems are applied to predict and simulate the performance of the components, and optimize the design of the components to meet specific functional and durability requirements.

[0023] S8. Ensure components are installed correctly.

[0024] During the on-site assembly phase, based on the design scheme in step S7, artificial intelligence guides the rapid and accurate installation of prefabricated components, ensuring the correct connection and fixation of each prefabricated component.

[0025] Optionally, S1 includes:

[0026] S11. Collect location information of the construction site and obtain coordinate data of the construction site through a GPS positioning system. The coordinate data includes X, Y, and Z coordinates.

[0027] X = Rcos(φ)cos(λ)

[0028] Y = Rcos(φ)sin(λ)

[0029] Z = Rsin(φ)

[0030] Where R is the Earth's radius, φ is latitude, and λ is longitude;

[0031] S12. Use a laser rangefinder and a level to measure the elevation of the construction site, and compile the measured absolute and relative elevations into elevation data:

[0032]

[0033] Where H is the elevation, D1 and D2 are the distances between the two points measured by the distance measuring instrument, and Δh i Let n be the elevation difference between each measuring point. (1 This represents the number of measurement points;

[0034] S13. Use flow velocity sensors and water level sensors to measure the water drop at the construction site and generate water drop data:

[0035]

[0036] Where Δh is the drop in water level, Q is the flow rate, A is the cross-sectional area, v is the flow velocity, g is the acceleration due to gravity, and h is the velocity. f For friction loss head;

[0037] S14. Obtain linear data of the length, width, and curvature of the culvert after assembly using a laser scanner or other measuring equipment:

[0038]

[0039] Where C is the curvature and L is the length of the culvert. The derivative of the curve;

[0040] S15. Calculate the safe lifting radius R'' based on the lifting weight W of the culvert:

[0041]

[0042] Where σ is the safety factor of the hoisting equipment, W is the hoisting weight, E is the elastic modulus of the material, and ν is the Poisson's ratio of the material;

[0043] S16. Based on the data collected in steps S11-S14, determine the specific coordinates of each hoisting position. Calculate and analyze the load-bearing capacity and stability of each position to avoid tilting or shifting during hoisting. The formulas for calculating the load-bearing capacity and stability are as follows:

[0044]

[0045] Where P is the load-bearing capacity, F is the applied force, A is the cross-sectional area, and M is the cross-sectional area. y and M z For bending moment, I y and I z Let y be the moment of inertia, and z be the coordinates of position.

[0046] S17. Combine the safe hoisting positions determined in steps S11-S16 with the construction schedule and resource allocation plan to construct an initial construction dataset. The initial construction dataset includes location data, elevation data, water level difference data, linear data, and safe hoisting radius data.

[0047]

[0048] Where D is the initial construction dataset, X i ,Y i Z i For location data, H i For elevation data, Δh i For the water drop data, C i For linear data, R' i 'For safe lifting radius data, n' (2) This represents the initial number of construction data points.

[0049] Optionally, S2 includes:

[0050] S21. Perform data cleaning, missing value imputation, and data standardization on the initial construction dataset D to obtain the preprocessed construction dataset.

[0051] S22. Using feature selection techniques, key features affecting construction performance are selected from the preprocessed construction dataset D to form a key feature dataset:

[0052] Using a hybrid feature selection algorithm based on information gain and mutual information, calculate F for each feature. i The overall score is calculated using the following formula:

[0053] Score(F i )=α·IG(F i )+β·MI(F i )

[0054] Among them, IG(F i ) is a feature F i Information gain, MI(F) i ) is a feature F i The mutual information, α and β are adjustment parameters used to balance the contributions of information gain and mutual information;

[0055] The top k features selected based on the comprehensive score are used as key features to form a key feature dataset:

[0056] F key ={F1,F2,…,F k}

[0057] S23, Based on key feature dataset F key An improved aggregation decision tree algorithm is applied for analysis:

[0058] Using multi-model ensemble techniques, decision trees, random forests, and gradient boosting trees are combined:

[0059]

[0060] in, For the final prediction result, h m (x) represents the prediction result of the decision tree model, f n (x) represents the prediction result of the random forest model, g p (x) represents the prediction result of the gradient boosting tree model, α m β n and γ p The weights are the corresponding model weights, and M, N, and P are the number of decision tree, random forest, and gradient boosting tree models, respectively.

[0061] In the process of multi-model ensemble, an adaptive weight adjustment mechanism is introduced to dynamically adjust the weights of each model by minimizing the prediction error, as shown in the following formula:

[0062]

[0063] Where, α m Let y be the weight of the m-th decision tree model. i x is the actual value.i Let λ be the input feature vector, λ be the adjustment coefficient, and Z be the normalization constant.

[0064] By employing feature selection techniques, the most important features are selected from the results of multi-model ensembles to construct an optimized feature set. The formula for constructing the optimized feature set is as follows:

[0065]

[0066] Among them, F opt For the optimized feature set, F i Let G be the candidate feature set, K be the number of folds in the cross-validation, and G be the number of folds in the cross-validation. k (F i H is the characteristic gain of the k-th fold. k (F i ) represents the feature loss at the k-th fold, and ∈ represents the smoothing term;

[0067] S24. Based on the analysis results of step S23, automatically identify the key performance indicators (KPIs) and potential efficiency bottlenecks in the construction process. i The key performance indicators include construction progress, material utilization rate, and equipment utilization rate. The efficiency bottlenecks include uneven allocation of construction resources and bottleneck links in the construction process. The calculation formulas for the key performance indicators are as follows:

[0068]

[0069] Wherein, KPI stands for Key Performance Indicator, ω i Let y be the weight of the i-th sample. i This is the actual value. For the predicted value, n (3 The number of samples;

[0070]

[0071] Among them, Bottleneck i For the i-th construction resource, Resource is the efficiency bottleneck indicator. i Let Capacity be the usage of the i-th resource. i For the total capacity of the i-th resource, Delay ij PlannedTime is the delay time of the j-th task on the i-th resource. ij Let n be the planned time for the j-th task. (4) The number of tasks.

[0072] Optionally, S3 includes:

[0073] S31. Based on the analysis results of step S2, construct an initial population. Each individual in the initial population represents a construction process and resource allocation scheme. The representation of the individual is as follows:

[0074]

[0075] Among them, P i Let n be the construction parameters for the i-th task. (5) Number of tasks;

[0076] S32. Define a fitness function f to evaluate the quality of each individual. The fitness function is calculated based on key performance indicators and efficiency bottlenecks. The formula for the fitness function is:

[0077]

[0078] Among them, KPI i For the key performance indicator of the i-th task, Bottleneck i The bottleneck for the i-th task is... and For the corresponding weights;

[0079] S33. Based on the fitness function f, select individuals with fitness higher than the set value as parents, and perform crossover to generate offspring individuals. The crossover operation adopts single-point crossover.

[0080]

[0081] in, This represents the i-th task parameter from the j-th parent;

[0082] S34. Introduce mutation operations into offspring individuals, adjusting some parameters of the individuals through random mutation and adaptive mutation rate. The formula for the mutation operation is:

[0083] P' i =P i +δ·randn(0,1)

[0084] Among them, P' i Here are the parameters after mutation, δ is the mutation amplitude, and randn(0,1) is a normally distributed random number with a mean of 0 and a variance of 1.

[0085] S35. Evaluate the quality of offspring individuals using the dynamic fitness function f, retain individuals with fitness values ​​higher than the set value to enter the next generation, and repeat steps S33 to S35 until the maximum number of iterations is reached or the fitness converges.

[0086] S36. The individual with the highest fitness is selected as the final construction process and resource allocation scheme, which includes the optimized configuration of manpower, machinery and equipment and materials.

[0087] Optionally, S4 includes:

[0088] S41. Convert the construction process and resource allocation plan generated in S3 into specific operation instructions, including the specific execution steps and required resource configurations for each construction task. The resources include manpower, machinery and equipment, and materials. The operation instructions are represented in the following form:

[0089] O = {T1,T2,…,T} m}

[0090] Among them, T i Here, m represents the operation instruction for the i-th construction task, and m is the number of tasks.

[0091] S42. The operation instructions are distributed to each construction task through the construction scheduling module, resource management module, and task allocation module of the artificial intelligence system. The specific steps include:

[0092] The construction scheduling module dynamically adjusts the execution order of operation instructions based on the real-time status and resource availability of the construction site to ensure the continuity and efficiency of the construction process.

[0093] The resource management module monitors the resource usage at the construction site, including manpower, machinery and equipment, and materials, updates the resource allocation plan in real time, and confirms the optimal resource configuration.

[0094] The task assignment module distributes operation instructions to specific construction personnel and equipment, and tracks the execution of tasks through the monitoring system to confirm the accurate implementation of the operation instructions.

[0095] Optionally, during construction, the artificial intelligence system monitors the environmental conditions, material status, and construction progress at the construction site in real time through a sensor network. The sensors include position sensors, elevation sensors, water drop sensors, and sensors monitoring the linearity of the assembled culvert. The data collected by these sensors forms a real-time monitoring dataset, represented as follows:

[0096]

[0097] Where, d i For the i-th real-time monitoring data, n (6) This refers to the number of monitoring data points.

[0098] Optionally, based on real-time monitoring dataset D real-time The artificial intelligence system assesses the status of the construction site in real time and dynamically adjusts operational instructions and resource allocation plans based on the assessment results.

[0099] When construction progress is detected to be lagging behind, the task allocation strategy is adjusted to increase manpower and equipment input;

[0100] When uneven resource usage is detected, resources are reallocated to ensure that the resource requirements of each task are met;

[0101] When changes in environmental conditions are detected, the construction process is adjusted to ensure construction safety and efficiency.

[0102] All adjusted operation instructions and resource allocation plans are recorded in the construction log, which includes task execution records, resource usage records, and adjustment records.

[0103] Optionally, S7 includes:

[0104] S71. During the design phase of prefabricated components, collect various parameters related to the prefabricated components, including material properties, geometric dimensions, and usage environment, and construct a prefabricated component design dataset, represented as:

[0105]

[0106] Among them, M i Let G be the material properties of the i-th prefabricated component. i E represents the geometric dimensions. i To use environmental parameters, n (7 The number of parameter data points;

[0107] S72. Applying an artificial intelligence system to the prefabricated component design dataset D design The analysis involves using machine learning models for performance prediction, including regression and neural network models. The performance prediction formula is as follows:

[0108]

[0109] Where σ is the activation function, These are the weight parameters of the neural network. The weights of the input parameters, Here, n1, n2, and n3 are the bias terms, and n1, n2, and n3 are the number of layers in the neural network.

[0110] S73. Based on the performance prediction results of step S72, the prefabricated component design is optimized using an optimization algorithm. The optimization objectives include meeting specific functional and durability requirements. The objective function of the optimization algorithm is:

[0111]

[0112] Among them, P target θ is the target performance value, β is the adjustment coefficient, and β is the target performance value. jThe weights for different features are γ, where γ is the bias term, and a j b j c j The weights of the input parameters;

[0113] S74. Verify the optimized prefabricated component design through simulation technology, including finite element analysis and computer simulation, to verify the performance of the prefabricated component in the actual use environment.

[0114] S75. Based on the results of simulation verification, further adjust and optimize the design of prefabricated components to ensure the functionality and durability requirements of prefabricated components in actual use;

[0115] S76. The final optimized prefabricated component design scheme is converted into manufacturing instructions to guide the production and manufacturing of prefabricated components. The manufacturing instructions are expressed in the following form:

[0116] O manufacturing ={C1,C2,…,C m‘’}

[0117] Among them, C i m'' represents the instruction for the i-th manufacturing step, and m'' represents the number of manufacturing steps.

[0118] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0119] (1) This invention utilizes an improved aggregation decision tree algorithm and a genetic algorithm to intelligently analyze and optimize the construction process, automatically identifying key performance indicators and potential efficiency bottlenecks in construction, and generating optimal construction processes and resource allocation schemes. This effectively reduces construction time and resource waste, and improves overall construction efficiency. The use of a dynamic fitness function and adaptive mutation rate makes the genetic algorithm more flexible and efficient in optimizing construction processes and resource allocation, adapting to real-time changes at the construction site and ensuring the optimality and practicality of the construction scheme.

[0120] (2) This invention uses the construction scheduling module, resource management module and task allocation module in the artificial intelligence system to monitor the resource usage at the construction site in real time, dynamically adjust the resource allocation plan, ensure the optimal allocation of resources, effectively reduce resource waste, improve the management efficiency of the construction site, and use the artificial intelligence system to generate specific operation instructions to ensure that the specific execution steps and required resource allocation of each construction task are clear and definite, thereby improving the continuity and execution efficiency of construction.

[0121] (3) During the construction process, the present invention monitors the environmental conditions, material status and construction progress of the construction site in real time through a sensor network, and constructs a dynamic construction dataset to ensure that all parameters of the construction site are under control. When an abnormal situation is detected, the artificial intelligence system can respond quickly and dynamically adjust the construction plan and operation instructions to ensure construction quality and safety. Attached Figure Description

[0122] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. In the drawings:

[0123] Figure 1 This is a schematic diagram of the process for an artificial intelligence-based prefabricated culvert on-site installation method proposed in this invention. Detailed Implementation

[0124] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0125] Reference Figure 1 The present invention provides an artificial intelligence-based prefabricated culvert on-site installation construction method, which includes the following steps;

[0126] S1. Collect initial construction data

[0127] Collect location, elevation, water drop, and linear data after the culvert is assembled; calculate the safe hoisting radius and hoisting position based on the hoisting weight; and construct the initial construction dataset.

[0128] S2. Analyze initial construction data

[0129] By applying an improved aggregated decision tree to analyze the initial construction dataset, key performance indicators and potential efficiency bottlenecks in the construction process are automatically identified.

[0130] S3. Generate the optimal construction process and resource allocation plan.

[0131] Based on the analysis results of step S2, an improved genetic algorithm is used to generate the optimal construction process and resource allocation scheme, where resources include manpower, machinery and equipment and materials.

[0132] S4. Convert and implement operation instructions

[0133] The construction process and resource allocation plan generated in step S3 are transformed into specific operation instructions, and the operation instructions are implemented through the artificial intelligence system to schedule relevant construction resources.

[0134] S5. Real-time monitoring of the construction site

[0135] During construction, a variety of sensors are used to monitor the environmental conditions, material status and worker behavior at the construction site in real time, and to build a dynamic construction dataset. The sensors include position sensors, elevation sensors, water drop sensors, sensors for monitoring the linearity of the culvert after assembly, and hoisting safety sensors.

[0136] S6. Dynamically adjust the construction plan.

[0137] Based on the dynamic construction dataset collected in real time in step S5, the construction plan and operation instructions are adjusted in real time through the artificial intelligence system to respond to any deviations or potential problems detected in real time.

[0138] S7, Optimize prefabricated component design

[0139] During the design phase of prefabricated components, artificial intelligence systems are applied to predict and simulate the performance of the components, and optimize the design of the components to meet specific functional and durability requirements.

[0140] S8. Ensure components are installed correctly.

[0141] During the on-site assembly phase, based on the design scheme in step S7, artificial intelligence guides the rapid and accurate installation of prefabricated components, ensuring the correct connection and fixation of each prefabricated component.

[0142] In this embodiment, S1 includes:

[0143] S11. Collect location information of the construction site, and obtain coordinate data of the construction site through a GPS positioning system. The coordinate data includes X, Y, and Z coordinates:

[0144] X = Rcos(φ)cos(λ)

[0145] Y = Rcos(φ)sin(λ)

[0146] Z = Rsin(φ)

[0147] Where R is the Earth's radius, φ is latitude, and λ is longitude;

[0148] S12. Use a laser rangefinder and a level to measure the elevation of the construction site, and compile the measured absolute and relative elevations into elevation data:

[0149]

[0150] Where H is the elevation, D1 and D2 are the distances between the two points measured by the distance measuring instrument, and Δh i Let n be the elevation difference between each measuring point. (1) This represents the number of measurement points;

[0151] S13. Use flow velocity sensors and water level sensors to measure the water drop at the construction site and generate water drop data:

[0152]

[0153] Where Δh is the drop in water level, Q is the flow rate, A is the cross-sectional area, v is the flow velocity, g is the acceleration due to gravity, and h is the velocity. f For friction loss head;

[0154] S14. Obtain linear data of the length, width, and curvature of the culvert after assembly using a laser scanner or other measuring equipment:

[0155]

[0156] Where C is the curvature and L is the length of the culvert. The derivative of the curve;

[0157] S15. Calculate the safe lifting radius R'' based on the lifting weight W of the culvert:

[0158]

[0159] Where σ is the safety factor of the hoisting equipment, W is the hoisting weight, E is the elastic modulus of the material, and ν is the Poisson's ratio of the material;

[0160] S16. Based on the data collected in steps S11-S14, determine the specific coordinates of each hoisting position. Calculate and analyze the load-bearing capacity and stability of each position to prevent tilting or displacement during hoisting. The formulas for calculating load-bearing capacity and stability are as follows:

[0161]

[0162] Where P is the load-bearing capacity, F is the applied force, A is the cross-sectional area, and M is the cross-sectional area. y and M z For bending moment, I y and I z Let y be the moment of inertia, and z be the coordinates of position.

[0163] S17. Combine the safe hoisting locations determined in steps S11-S16 with the construction schedule and resource allocation plan to construct an initial construction dataset. The initial construction dataset includes location data, elevation data, water level difference data, linear data, and safe hoisting radius data.

[0164]

[0165] Where D is the initial construction dataset, X i ,Y i Zi For location data, H i For elevation data, Δh i For the water drop data, C i For linear data, R'' i For safe lifting radius data, n (2) This represents the initial number of construction data points.

[0166] In this embodiment, S2 includes:

[0167] S21. Perform data cleaning, missing value imputation, and data standardization on the initial construction dataset D to obtain the preprocessed construction dataset.

[0168] S22. Using feature selection techniques, key features affecting construction performance are selected from the preprocessed construction dataset D to form a key feature dataset:

[0169] Using a hybrid feature selection algorithm based on information gain and mutual information, calculate F for each feature. i The overall score is calculated using the following formula:

[0170] Score(F i )=α·IG(F i )+β·MI(F i )

[0171] Among them, IG(F i ) is a feature F i Information gain, MI(F) i ) is a feature F i The mutual information, α and β are adjustment parameters used to balance the contributions of information gain and mutual information;

[0172] The top k features selected based on the comprehensive score are used as key features to form a key feature dataset:

[0173] F key ={F1,F2,…,F k}

[0174] S23, Based on key feature dataset F key An improved aggregation decision tree algorithm is applied for analysis:

[0175] Using multi-model ensemble techniques, decision trees, random forests, and gradient boosting trees are combined:

[0176]

[0177] in, For the final prediction result, h m (x) represents the prediction result of the decision tree model, f n(x) represents the prediction result of the random forest model, g p (x) represents the prediction result of the gradient boosting tree model, α m β n and γ p The weights are the corresponding model weights, and M, N, and P are the number of decision tree, random forest, and gradient boosting tree models, respectively.

[0178] In the process of multi-model ensemble, an adaptive weight adjustment mechanism is introduced to dynamically adjust the weights of each model by minimizing the prediction error, as shown in the following formula:

[0179]

[0180] Where, α m Let y be the weight of the m-th decision tree model. i x is the actual value. i Let λ be the input feature vector, Z be the adjustment coefficient, and β be the normalization constant. n and γ p The adjustment principle is the same;

[0181] By employing feature selection techniques, the most important features are chosen from the results of multi-model ensembles to construct an optimized feature set. The formula for constructing the optimized feature set is as follows:

[0182]

[0183] Among them, F opt For the optimized feature set, F i Let G be the candidate feature set, K be the number of folds in the cross-validation, and G be the number of folds in the cross-validation. k (F i H is the characteristic gain of the k-th fold. k (F i ) represents the feature loss at the k-th fold, and ∈ represents the smoothing term;

[0184] S24. Based on the analysis results of step S23, automatically identify the key performance indicators (KPIs) and potential efficiency bottlenecks in the construction process. i Key performance indicators include construction progress, material utilization rate, and equipment utilization rate. Efficiency bottlenecks include uneven allocation of construction resources and bottleneck links in the construction process. The calculation formulas for key performance indicators are as follows:

[0185]

[0186] Wherein, KPI stands for Key Performance Indicator, ω i Let y be the weight of the i-th sample. i This is the actual value. For the predicted value, n (3) The number of samples;

[0187]

[0188] Among them, Bottleneck i For the i-th construction resource, Resource is the efficiency bottleneck indicator. i Let Capacity be the usage of the i-th resource. i For the total capacity of the i-th resource, Delay ij PlannedTime is the delay time of the j-th task on the i-th resource. ij Let n be the planned time for the j-th task. (4) The number of tasks.

[0189] In this embodiment, S3 includes:

[0190] S31. Based on the analysis results of step S2, construct an initial population. Each individual in the initial population represents a construction process and resource allocation scheme. The representation of the individual is as follows:

[0191]

[0192] Among them, P i Let n be the construction parameters for the i-th task. (5) Number of tasks;

[0193] S32. Define a fitness function f to evaluate the performance of each individual. The fitness function is calculated based on key performance indicators and efficiency bottlenecks. The formula for the fitness function is:

[0194]

[0195] Among them, KPI i For the key performance indicator of the i-th task, Bottleneck i The bottleneck for the i-th task is... and For the corresponding weights;

[0196] S33. Based on the fitness function f, select individuals with fitness higher than the set value as parents, and perform crossover to generate offspring individuals. The crossover operation uses single-point crossover.

[0197]

[0198] in, This represents the i-th task parameter from the j-th parent;

[0199] S34. Introduce mutation operations into offspring individuals, adjusting some parameters of the individuals through random mutation and adaptive mutation rate. The formula for the mutation operation is:

[0200] P' i =P i +δ·randn(0,1)

[0201] Among them, P' i Here are the parameters after mutation, δ is the mutation amplitude, and randn(0,1) is a normally distributed random number with a mean of 0 and a variance of 1.

[0202] S35. Evaluate the quality of offspring individuals using the dynamic fitness function f, retain individuals with fitness values ​​higher than the set value to enter the next generation, and repeat steps S33 to S35 until the maximum number of iterations is reached or the fitness converges.

[0203] S36. The individual with the highest fitness is used as the final construction process and resource allocation plan. The final plan includes the optimized allocation of manpower, machinery and equipment and materials.

[0204] In this embodiment, S4 includes:

[0205] S41. Transform the construction process and resource allocation plan generated in S3 into specific operation instructions, including the specific execution steps of each construction task and the required resource allocation. Resources include manpower, machinery and equipment, and materials. The operation instructions are represented in the following form:

[0206] O = {T1, T2, ..., Tm}

[0207] Among them, T i Here, m represents the operation instruction for the i-th construction task, and m is the number of tasks.

[0208] S42. The operation instructions are distributed to each construction task through the construction scheduling module, resource management module, and task allocation module of the artificial intelligence system. The specific steps include:

[0209] The construction scheduling module dynamically adjusts the execution order of operation instructions based on the real-time status and resource availability of the construction site to ensure the continuity and efficiency of the construction process.

[0210] The resource management module monitors the resource usage at the construction site, including manpower, machinery and equipment, and materials, updates the resource allocation plan in real time, and confirms the optimal resource configuration.

[0211] The task assignment module distributes operation instructions to specific construction personnel and equipment, and tracks the execution of tasks through the monitoring system to confirm the accurate implementation of the operation instructions.

[0212] In this embodiment, during construction, the artificial intelligence system monitors the environmental conditions, material status, and construction progress at the construction site in real time through a sensor network. The sensors include position sensors, elevation sensors, water drop sensors, and sensors monitoring the linearity of the assembled culvert. The data collected by the sensors forms a real-time monitoring dataset, represented as follows:

[0213]

[0214] Where, d i For the i-th real-time monitoring data, n (6) This refers to the number of monitoring data points.

[0215] In this embodiment, based on the real-time monitoring dataset D real-time The artificial intelligence system assesses the status of the construction site in real time and dynamically adjusts operational instructions and resource allocation plans based on the assessment results.

[0216] When construction progress is detected to be lagging behind, the task allocation strategy is adjusted to increase manpower and equipment input;

[0217] When uneven resource usage is detected, resources are reallocated to ensure that the resource requirements of each task are met;

[0218] When changes in environmental conditions are detected, the construction process is adjusted to ensure construction safety and efficiency.

[0219] All adjusted operation instructions and resource allocation plans are recorded in the construction log, which includes task execution records, resource usage records, and adjustment records.

[0220] In this embodiment, S7 includes:

[0221] S71. During the design phase of prefabricated components, collect various parameters related to the prefabricated components, including material properties, geometric dimensions, and usage environment, and construct a prefabricated component design dataset, represented as:

[0222]

[0223] Among them, M i Let G be the material properties of the i-th prefabricated component. i E represents the geometric dimensions. i To use environmental parameters, n (7) The number of parameter data points;

[0224] S72. Applying an artificial intelligence system to the prefabricated component design dataset D design The analysis utilizes machine learning models for performance prediction, including regression and neural network models. The performance prediction formula is as follows:

[0225]

[0226] Where σ is the activation function, These are the weight parameters of the neural network. The weights of the input parameters, Here, n1, n2, and n3 are the bias terms, and n1, n2, and n3 are the number of layers in the neural network.

[0227] S73. Based on the performance prediction results of step S72, the prefabricated component design is optimized using an optimization algorithm. The optimization objectives include meeting specific functional and durability requirements. The objective function of the optimization algorithm is:

[0228]

[0229] Among them, P target θ is the target performance value, β is the adjustment coefficient, and β is the target performance value. j The weights for different features are γ, where γ is the bias term, and a j b j c j The weights of the input parameters;

[0230] S74. Verify the optimized prefabricated component design through simulation technology, including finite element analysis and computer simulation, to verify the performance of the prefabricated components in actual use environments.

[0231] S75. Based on the results of simulation verification, further adjust and optimize the design of prefabricated components to ensure the functionality and durability requirements of prefabricated components in actual use;

[0232] S76. The final optimized prefabricated component design scheme is transformed into a manufacturing instruction to guide the production and manufacturing of the prefabricated components. The manufacturing instruction is represented in the following form:

[0233] O manufacturing ={C1,C2,…,C m‘’}

[0234] Among them, C i Let m'' be the instruction for the i-th manufacturing step, and m'' be the number of manufacturing steps.

[0235] This invention utilizes an improved aggregation decision tree algorithm and a genetic algorithm to intelligently analyze and optimize the construction process. It automatically identifies key performance indicators and potential efficiency bottlenecks during construction, generating optimal construction processes and resource allocation schemes. This effectively reduces construction time and resource waste, improving overall construction efficiency. By employing a dynamic fitness function and adaptive mutation rate, the genetic algorithm becomes more flexible and efficient in optimizing construction processes and resource allocation, adapting to real-time changes at the construction site and ensuring the optimality and practicality of the construction plan.

[0236] This invention utilizes a construction scheduling module, a resource management module, and a task allocation module within an artificial intelligence system to monitor resource usage at the construction site in real time, dynamically adjust resource allocation schemes, ensure optimal resource allocation, effectively reduce resource waste, and improve the management efficiency of the construction site. By generating specific operation instructions using the artificial intelligence system, the specific execution steps and required resource allocation for each construction task are clearly defined, thereby improving the continuity and efficiency of construction.

[0237] During construction, this invention uses a sensor network to monitor the environmental conditions, material status, and construction progress of the construction site in real time, constructing a dynamic construction dataset to ensure that all parameters at the construction site are under control. When anomalies are detected, the artificial intelligence system can respond quickly and dynamically adjust the construction plan and operation instructions to ensure construction quality and safety.

[0238] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A prefabricated culvert installation method based on artificial intelligence, characterized in that, Includes the following steps; S1. Collect initial construction data Collect location, elevation, water drop, and linear data after the culvert is assembled. Calculate the safe hoisting radius and hoisting position based on the hoisting weight to construct the initial construction dataset. S2. Analyze initial construction data By applying an improved aggregated decision tree to analyze the initial construction dataset, key performance indicators and potential efficiency bottlenecks in the construction process are automatically identified. S3. Generate the optimal construction process and resource allocation plan. Based on the analysis results of step S2, an improved genetic algorithm is used to generate the optimal construction process and resource allocation scheme, where resources include manpower, machinery and equipment and materials. S4. Convert and implement operation instructions The construction process and resource allocation plan generated in step S3 are transformed into specific operation instructions, and the operation instructions are implemented through the artificial intelligence system to schedule relevant construction resources. S5. Real-time monitoring of the construction site During construction, a variety of sensors are used to monitor the environmental conditions, material status and worker behavior at the construction site in real time, and to build a dynamic construction dataset. The sensors include position sensors, elevation sensors, water drop sensors, sensors for monitoring the linearity of the culvert after assembly, and hoisting safety sensors. S6. Dynamically adjust the construction plan. Based on the dynamic construction dataset collected in real time in step S5, the construction plan and operation instructions are adjusted in real time through the artificial intelligence system to respond to any deviations or potential problems detected in real time. S7. Optimize prefabricated component design During the design phase of prefabricated components, artificial intelligence systems are applied to predict and simulate the performance of the components, and optimize the design of the components to meet specific functional and durability requirements. S8. Ensure components are installed correctly. During the on-site assembly phase, based on the design scheme in step S7, artificial intelligence guides the rapid and accurate installation of prefabricated components, ensuring the correct connection and fixation of each prefabricated component.

2. The prefabricated culvert installation method based on artificial intelligence according to claim 1, characterized in that, S1 includes: S11. Collect location information of the construction site, and obtain coordinate data of the construction site through a GPS positioning system. The coordinate data includes X, Y, and Z coordinates: ； in, For the Earth's radius, Latitude Longitude; S12. Use a laser rangefinder and level to measure the elevation of the construction site, and compile the measured absolute and relative elevations into elevation data: ； in, For elevation, and The distance between two points measured by the rangefinder. The elevation difference between each measuring point, This represents the number of measurement points; S13. Use flow velocity sensors and water level sensors to measure the water drop at the construction site and generate water drop data: ； in, For the drop in water level, For traffic, For cross-sectional area, For flow rate, It is the acceleration due to gravity. For friction loss head; S14. Obtain linear data of the length, width, and curvature of the culvert after assembly using a laser scanner or other measuring equipment: ； in, For curvature, The length of the culvert The derivative of the curve; S15. Based on the lifting weight of the culvert Calculate the safe lifting radius : ； in, To ensure the safety factor of the hoisting equipment, For lifting weight, The elastic modulus of the material, The Poisson's ratio of the material; S16. Based on the data collected in steps S11-S14, determine the specific coordinates of each hoisting position. Calculate and analyze the load-bearing capacity and stability of each position to avoid tilting or shifting during hoisting. The formulas for calculating the load-bearing capacity and stability are as follows: ； in, For load-bearing capacity, For the action force, For cross-sectional area, and For bending moment, and Let y be the moment of inertia, and z be the coordinates of position. S17. Combine the safe hoisting positions determined in steps S11-S16 with the construction schedule and resource allocation plan to construct an initial construction dataset. The initial construction dataset includes location data, elevation data, water level difference data, linear data, and safe hoisting radius data. ； in, For the initial construction dataset, For location data, For elevation data, For water drop data, For linear data, For safe lifting radius data, This represents the initial number of construction data points.

3. The method for on-site installation and construction of prefabricated culverts based on artificial intelligence according to claim 2, characterized in that, S2 includes: S21. Initial construction dataset Data cleaning, missing value imputation, and data standardization are performed to obtain a preprocessed construction dataset. S22. Using feature selection techniques, from the preprocessed construction dataset... Select the key features that affect construction performance to form a key feature dataset: A hybrid feature selection algorithm based on information gain and mutual information is used to calculate each feature. The overall score is calculated using the following formula: ； in, Features Information gain Features mutual information, and The adjustment parameters are used to balance the contributions of information gain and mutual information. Selected based on comprehensive score These features are used as key features to form a key feature dataset: ； S23, Based on key feature datasets An improved aggregation decision tree algorithm is applied for analysis: Using multi-model ensemble techniques, decision trees, random forests, and gradient boosting trees are combined: ； in, For the final prediction result, The prediction results of the decision tree model, These are the prediction results from the random forest model. The prediction results of the gradient boosting tree model, , and For the weights of the corresponding model, , and These represent the number of decision tree, random forest, and gradient boosting tree models, respectively. In the process of multi-model ensemble, an adaptive weight adjustment mechanism is introduced to dynamically adjust the weights of each model by minimizing the prediction error, as shown in the following formula: ； in, For the first The weights of a decision tree model, This is the actual value. For the input feature vector, To adjust the coefficient, This is a normalization constant; By employing feature selection techniques, the most important features are selected from the results of multi-model ensembles to construct an optimized feature set. The formula for constructing the optimized feature set is as follows: ； in, For the optimized feature set, For candidate feature set, The number of folds for cross-validation. For the first The characteristic gain of the fold, For the first Feature loss of folds, For smoothing terms; S24. Based on the analysis results of step S23, automatically identify the key performance indicators in the construction process. and potential efficiency bottlenecks The key performance indicators include construction progress, material utilization rate, and equipment utilization rate. The efficiency bottlenecks include uneven allocation of construction resources and bottleneck links in the construction process. The calculation formulas for the key performance indicators are as follows: ； in, As a key performance indicator, For the first The weights of each sample, This is the actual value. For predicted values, The number of samples; ； in, For the first Efficiency bottleneck indicators of construction resources For the first The usage of each resource For the first The total capacity of each resource For the first The task in the first The latency of each resource For the first The planned time for each task The number of tasks.

4. The prefabricated culvert installation method based on artificial intelligence according to claim 3, characterized in that, S3 includes: S31. Based on the analysis results of step S2, construct an initial population. Each individual in the initial population represents a construction process and resource allocation scheme. The representation of the individual is as follows: ； in, For the first Construction parameters for each task, Number of tasks; S32. Define the fitness function The fitness function is used to evaluate the performance of each individual. It is calculated based on key performance indicators and efficiency bottlenecks, and the formula for the fitness function is: ； in, For the first Key performance indicators for each task For the first The efficiency bottleneck of each task and For the corresponding weights; S33. Based on the fitness function Individuals with fitness higher than a set value are selected as parents, and crossover is performed to generate offspring. The crossover operation uses a single-point crossover. ； in, Indicates from the The first of the fathers Task parameters; S34. Introduce mutation operations into offspring individuals, adjusting some parameters of the individuals through random mutation and adaptive mutation rate. The formula for the mutation operation is: ； in, These are the mutated parameters. The range of variation, These are normally distributed random numbers with a mean of 0 and a variance of 1. S35, Through dynamic fitness function Evaluate the quality of offspring individuals, retain individuals with fitness higher than the set value to enter the next generation, and repeat steps S33 to S35 until the maximum number of iterations is reached or fitness converges; S36. The individual with the highest fitness is selected as the final construction process and resource allocation plan, which includes the optimal allocation of manpower, machinery and equipment and materials.

5. The method for on-site installation of prefabricated culverts based on artificial intelligence according to claim 4, characterized in that, S4 includes: S41. Convert the construction process and resource allocation plan generated in S3 into specific operation instructions, including the specific execution steps and required resource configurations for each construction task. The resources include manpower, machinery and equipment, and materials. The operation instructions are represented in the following form: ； in, For the first Operational instructions for each construction task. Number of tasks; S42. The operation instructions are distributed to each construction task through the construction scheduling module, resource management module, and task allocation module of the artificial intelligence system. The specific steps include: The construction scheduling module dynamically adjusts the execution order of operation instructions based on the real-time status and resource availability of the construction site to ensure the continuity and efficiency of the construction process. The resource management module monitors the resource usage at the construction site, including manpower, machinery and equipment, and materials, updates the resource allocation plan in real time, and confirms the optimal resource configuration. The task assignment module distributes operation instructions to specific construction personnel and equipment, and tracks the execution of tasks through the monitoring system to confirm the accurate implementation of the operation instructions.

6. The method for on-site installation of prefabricated culverts based on artificial intelligence according to claim 5, characterized in that, During construction, the artificial intelligence system monitors the environmental conditions, material status, and construction progress at the construction site in real time through a sensor network. The sensors include position sensors, elevation sensors, water drop sensors, and sensors monitoring the linearity of the assembled culvert. The data collected by these sensors forms a real-time monitoring dataset, represented as follows: ； in, For the first Real-time monitoring data, This refers to the number of monitoring data points.

7. The prefabricated culvert installation method based on artificial intelligence according to claim 6, characterized in that, Based on real-time monitoring dataset The artificial intelligence system assesses the status of the construction site in real time and dynamically adjusts operational instructions and resource allocation plans based on the assessment results. When construction progress is detected to be lagging behind, the task allocation strategy is adjusted to increase manpower and equipment input; When uneven resource usage is detected, resources are reallocated to ensure that the resource requirements of each task are met; When changes in environmental conditions are detected, the construction process is adjusted to ensure construction safety and efficiency. All adjusted operation instructions and resource allocation plans are recorded in the construction log, which includes task execution records, resource usage records, and adjustment records.

8. The method for on-site installation of prefabricated culverts based on artificial intelligence according to claim 7, characterized in that, S7 includes: S71. During the design phase of prefabricated components, collect various parameters related to the prefabricated components, including material properties, geometric dimensions, and usage environment, and construct a prefabricated component design dataset, represented as: ； in, For the first Material properties of each prefabricated component Geometric dimensions To use environmental parameters, The number of parameter data points; S72. Applying artificial intelligence systems to prefabricated component design datasets The analysis involves using machine learning models for performance prediction, including regression and neural network models. The performance prediction formula is as follows: ； in, For activation function, These are the weight parameters of the neural network. The weights of the input parameters, For bias terms, The number of layers in the neural network; S73. Based on the performance prediction results of step S72, the prefabricated component design is optimized using an optimization algorithm. The optimization objectives include meeting specific functional and durability requirements. The objective function of the optimization algorithm is: ； in, For the target performance value, For adjustment coefficients, Weights for different features, For bias terms, , , The weights of the input parameters; S74. Verify the optimized prefabricated component design through simulation technology, including finite element analysis and computer simulation, to verify the performance of the prefabricated component in the actual use environment. S75. Based on the results of simulation verification, further adjust and optimize the design of prefabricated components to ensure the functionality and durability requirements of prefabricated components in actual use; S76. The final optimized prefabricated component design scheme is converted into manufacturing instructions to guide the production and manufacturing of prefabricated components. The manufacturing instructions are expressed in the following form: ； in, For the first Instructions for each manufacturing step, The number of manufacturing steps.

Images