Construction engineering risk assessment method based on artificial intelligence
By using AI-based data processing and model building, the problems of low efficiency and poor real-time performance in traditional construction engineering risk assessment have been solved, enabling efficient and personalized risk assessment and real-time monitoring, and optimizing resource allocation.
Patent Information
- Application Number
- CN202411414038.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-11
- Publication Date
- 2025-01-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional risk assessment methods for construction projects rely on historical data and expert judgment, which are inefficient, prone to errors, and difficult to achieve real-time risk assessment and monitoring.
An AI-based approach was adopted, including modules for data collection and preprocessing, feature selection, model building, and training. A linear support vector machine model was used for risk assessment, and the model was evaluated based on accuracy, recall, and F1 score.
It improves the accuracy and efficiency of risk prediction in construction projects, enables real-time monitoring, reduces human error, and provides personalized risk assessment and optimized resource allocation.
Smart Images
Figure CN119294817A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field, and in particular to a construction engineering risk assessment method based on artificial intelligence. Background Art
[0002] Through risk assessment, we can identify and prevent potential risks that may lead to casualties, ensure the safety of construction site workers, and promptly discover and deal with safety hazards. This can significantly reduce the probability of accidents during the construction process, protect the lives of workers, and help identify risk factors that may lead to property losses in advance, and take appropriate measures to avoid or reduce losses. Construction project risk assessment is an important means to ensure the smooth progress of construction projects, protect the safety of personnel and property, maintain social stability, and promote technological progress. It plays an irreplaceable role in the development of the construction industry.
[0003] Many risk assessment methods rely on historical data to predict future risks. Expert judgment often plays a key role in risk assessment. However, expert judgment may be affected by personal experience, bias, or incomplete available information. Traditional methods usually rely on manual data processing, which is inefficient and prone to errors, making it difficult to achieve real-time risk assessment and monitoring.
[0004] To this end, we propose an artificial intelligence-based construction project risk assessment method to solve the existing problems. Summary of the invention
[0005] The purpose of the present invention is to propose a construction project risk assessment method based on artificial intelligence in response to the problems existing in the background technology.
[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a construction project risk assessment method based on artificial intelligence, comprising a data collection and preprocessing module, a feature selection module, a model building and training module and a model evaluation module, wherein the data collection and preprocessing module processes data related to the construction project, the feature selection module selects relevant feature data from the processed data set and inputs it into the model building and training module for training and optimizing model parameters, and the performance of the model is evaluated by the model evaluation module using accuracy, recall rate and F1 score indicators.
[0007] Preferably, the data collection and preprocessing module includes a data collection module and a data preprocessing module. The data collection module includes determining data requirements, designing a data collection plan, implementing data collection, data storage and management, data quality checking and data integration. Through these detailed data collection steps, it can be ensured that high-quality data is collected, providing a solid foundation for construction project risk assessment.
[0008] Preferably, the data preprocessing module includes data cleaning, denoising and normalization. The data cleaning uses a deduplication function in a database or programming language to ensure that all duplicate data rows are deleted and unique data is retained. The denoising uses filters or smoothing techniques (such as moving average, median filtering) to reduce random noise in the data. For time series data, a differential method can be used to eliminate trend and seasonal effects. The normalization converts the data into a distribution with zero mean and unit variance, scales the data to a specific range, and scales different features to ensure that they have the same importance in model training. Through these refined steps, the quality of the data can be effectively improved, laying a solid foundation for subsequent data analysis and modeling.
[0009] Preferably, the feature selection module includes feature recognition, feature extraction, feature analysis, feature selection algorithm and feature selection result verification. Through these steps, the feature selection process can be refined to ensure that the final selected feature set can effectively predict construction project risks.
[0010] Preferably, the feature selection algorithm adopts filtering selection, which selects features based on univariate statistical tests, and the univariate statistical test adopts ANOVA, i.e., analysis of variance, and the ANOVA includes proposing hypotheses: null hypothesis (H0): there is no significant difference in the means between the groups, alternative hypothesis (Ha): the mean of at least one group is significantly different from the means of other groups;
[0011] Collect data: determine the experimental design, collect data from each group, and ensure that the data meet the assumptions of ANOVA, such as normality and homogeneity of variance;
[0012] Calculate the total sum of squares (SST): The total sum of squares is the sum of the squares of the differences between all observations and their grand mean, which represents the total variation of the data;
[0013] Calculate the between-group sum of squares (SSB): The between-group sum of squares is the sum of the squares of the differences between the means of each group and its grand mean, multiplied by the sample size of each group, which represents the variation caused by the differences in means between different groups;
[0014] Calculate the within-group sum of squares (SSW): The within-group sum of squares is the sum of the squares of the differences between the observations within each group and their group means, which represents the variation within each group;
[0015] Calculate the mean square (MS): The between-group mean square (MSB) is the between-group sum of squares divided by the between-group degrees of freedom. The within-group mean square (MSW) is the within-group sum of squares divided by the within-group degrees of freedom;
[0016] Calculate the F value: The F value is the ratio of the mean square between groups to the mean square within groups, that is, F = MSB / MSW;
[0017] Determine the P value: Use an F distribution table or statistical software to determine the P value of the F value. The P value represents the probability of observing the current F value or a more extreme situation when the null hypothesis is true.
[0018] Make a decision: If the P value is less than the significance level (usually 0.05), the null hypothesis is rejected, and it is believed that the mean of at least one group is significantly different from the means of other groups. If the P value is greater than the significance level, the null hypothesis cannot be rejected, and it is believed that there is no significant difference in the means between the groups.
[0019] Preferably, the model construction and training module includes model construction and model training. The model construction adopts a linear support vector machine model, which performs well when processing high-dimensional data and can effectively classify risks. The steps of model training include function interval, geometric interval and interval maximization.
[0020] Preferably, the model evaluation module includes evaluation indicators, cross-validation and hyperparameter tuning, and the evaluation indicators include accuracy, precision, recall and F1 score. Through the above-mentioned refinement steps, the linear support vector machine model can be comprehensively evaluated, so as to select the model configuration that best suits the specific problem.
[0021] Compared with the prior art, the present invention has the following beneficial effects:
[0022] During use of the present invention;
[0023] Improve prediction accuracy: Construction project risk assessment requires processing a large amount of complex data. Artificial intelligence technology can improve the prediction accuracy of potential risks through deep learning and pattern recognition, reduce misjudgments and missed judgments, and thus provide more reliable decision support for projects;
[0024] Enhanced data processing capabilities: Construction engineering data usually includes multi-source heterogeneous data such as structural design, construction process, geographic information, etc. Artificial intelligence methods can efficiently process this data, including unstructured data, and improve the efficiency and comprehensiveness of risk assessment;
[0025] Real-time monitoring: In construction projects, risks are changing dynamically. Artificial intelligence technology can achieve real-time monitoring of project risks, timely detect and warn of potential risks, and provide instant information for project management and emergency response;
[0026] Facilitate personalized assessment: Different construction projects have their own specific risk characteristics. AI-based methods can customize risk assessment models according to the specific circumstances of the project, achieving more personalized risk assessment;
[0027] Optimize resource allocation: Risk assessment through artificial intelligence can help project teams more accurately identify high-risk areas, thereby rationally allocating resources and optimizing risk management strategies;
[0028] Reduce human error: Traditional risk assessment methods rely heavily on the subjective judgment of experts and are susceptible to human factors. AI methods can reduce this subjectivity and reduce the risk of human error.
[0029] To sum up, the proposal of the construction project risk assessment method based on artificial intelligence is of great significance and value in improving the safety level of construction projects, optimizing project management, reducing economic losses, and protecting personnel safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION
[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.
[0032] Embodiment 1
[0033] like Figure 1 As shown, the present invention proposes an artificial intelligence-based construction project risk assessment method, which includes a data collection and preprocessing module, a feature selection module, a model building and training module and a model evaluation module. The data collection and preprocessing module processes data related to the construction project, the feature selection module selects relevant feature data from the processed data set and inputs it into the model building and training module for training and optimizing model parameters, and the performance of the model is evaluated by the model evaluation module using accuracy, recall rate and F1 score indicators.
[0034] Embodiment 2
[0035] like Figure 1 As shown, the invention proposes a construction engineering risk assessment method based on artificial intelligence. Compared with the first embodiment, this embodiment further includes: the data collection and preprocessing module includes a data collection module and a data preprocessing module. The data collection module includes determining data requirements, designing a data collection plan, implementing data collection, data storage and management, data quality inspection and data integration. Determining data requirements includes clarifying which data needs to be collected and determining the source of the data, such as project documents, construction logs, historical accident reports, environmental monitoring data, etc., according to the features determined in the feature selection stage;
[0036] Designing a data collection plan includes developing a detailed data collection plan, including the type, format, frequency of collection, and responsible persons of the data;
[0037] Data collection implementation collects data through on-site surveys, questionnaires, sensor data collection, etc. to ensure the accuracy and completeness of the data and avoid missing or incorrect data;
[0038] Data integration integrates data from different sources into a unified format to facilitate subsequent analysis and processing, ensuring that the integrated data set can reflect all aspects of the construction project. Through these detailed data collection steps, high-quality data can be collected, providing a solid foundation for construction project risk assessment.
[0039] The data preprocessing module includes data cleaning, denoising and normalization. Data cleaning uses the deduplication function in the database or programming language to determine which columns (features) are used to judge duplication. Usually, a unique identification column is selected based on business logic to decide whether to retain certain records in the duplicate items, such as retaining the first or last record. According to the size and structure of the data set, a suitable deduplication method is selected and the deduplication function is used to perform the deduplication operation to ensure that the deduplication data set retains the most important business information. The deduplication function can use DISTINCT in SQL to ensure that all duplicate data rows are deleted and unique data is retained.
[0040] Denoising uses filters or smoothing techniques (such as moving average, median filtering) to reduce random noise in the data. For time series data, you can use difference methods to remove trends and seasonal effects. You can use tools such as Matplotlib and Seaborn to identify outliers and noise in the data.
[0041] Normalization converts the data into a distribution with zero mean and unit variance, scales the data to a specific range, and scales different features to ensure that they have the same importance in model training. Through these refined steps, the quality of the data can be effectively improved, laying a solid foundation for subsequent data analysis and modeling.
[0042] The feature selection module includes feature identification, feature extraction, feature analysis, feature selection algorithm and feature selection result verification. Through these steps, the feature selection process can be refined to ensure that the final selected feature set can effectively predict the risks of construction projects. Feature identification determines the project scale, project type, budget, project cycle, geographical location, project history (such as previous delays or cost overruns), construction methods, construction technology, construction team experience, construction equipment, construction progress, construction quality, climate conditions, geological conditions, surrounding environment, traffic conditions and risk records of similar projects, accident frequency, historical costs and progress deviations. The above features are extracted from the original data. For non-numerical features, such as project type or construction method, encoding conversion may be required, and then statistical methods are used to determine the relationship between features and risks. Visualization tools (such as scatter plots and box plots) are used to observe feature distribution and potential relationship with risks.
[0043] The feature selection algorithm uses filtering selection, which selects features based on univariate statistical tests. The univariate statistical test uses ANOVA, or analysis of variance. ANOVA includes the following hypotheses: Null hypothesis (H0): There is no significant difference in the means between the groups; Alternative hypothesis (Ha): There is a significant difference between the means of at least one group and the means of the other groups;
[0044] Collect data: determine the experimental design, collect data from each group, and ensure that the data meet the assumptions of ANOVA, such as normality and homogeneity of variance;
[0045] Calculate the total sum of squares (SST): The total sum of squares is the sum of the squares of the differences between all observations and their grand mean, which represents the total variation of the data;
[0046] Calculate the between-group sum of squares (SSB): The between-group sum of squares is the sum of the squares of the differences between the means of each group and its grand mean, multiplied by the sample size of each group, which represents the variation caused by the differences in means between different groups;
[0047] Calculate the within-group sum of squares (SSW): The within-group sum of squares is the sum of the squares of the differences between the observations within each group and their group means, which represents the variation within each group;
[0048] Calculate the mean square (MS): The between-group mean square (MSB) is the between-group sum of squares divided by the between-group degrees of freedom. The within-group mean square (MSW) is the within-group sum of squares divided by the within-group degrees of freedom;
[0049] Calculate the F value: The F value is the ratio of the mean square between groups to the mean square within groups, that is, F = MSB / MSW;
[0050] Determine the P value: Use an F distribution table or statistical software to determine the P value of the F value. The P value represents the probability of observing the current F value or a more extreme situation when the null hypothesis is true.
[0051] Make a decision: If the P value is less than the significance level (usually 0.05), the null hypothesis is rejected, and it is believed that the mean of at least one group is significantly different from the means of other groups. If the P value is greater than the significance level, the null hypothesis cannot be rejected, and it is believed that there is no significant difference in the means between the groups. Finally, cross-validation or leave-one-out method is used to verify the results of feature selection to ensure that the selected features can show consistency on different data subsets.
[0052] Embodiment 3
[0053] like Figure 1 As shown, the invention proposes a construction engineering risk assessment method based on artificial intelligence. Compared with the first embodiment, this embodiment further includes: the model construction and training module includes model construction and model training. The model construction adopts a linear support vector machine model, which performs well in processing high-dimensional data and can effectively classify risks. The model training step includes the function interval:
[0054] On the training data set, the function interval is expressed as:
[0055]
[0056] in is the category label (valued as +1 or -1), w is the weight vector, b is the bias term, and x is the feature vector.
[0057] The geometric interval function interval takes into account the result after ||w||:
[0058]
[0059] The maximum margin is expressed as:
[0060] The goal of SVM is to find the hyperplane that maximizes the margin, that is:
[0061]
[0062] is equivalent to:
[0063]
[0064] So that for all training samples (x i ,y i )satisfy:
[0065] y i (w·x i +b)≥1.
[0066] The model evaluation module includes evaluation indicators, cross-validation and hyperparameter tuning. The evaluation indicators include accuracy, precision, recall and F1 score. Accuracy is the most intuitive evaluation indicator. The calculation formula is:
[0067]
[0068] Precision refers to the proportion of samples that are actually positive among all samples predicted to be positive. The calculation formula is:
[0069]
[0070] Among them, TP is a true positive example and FP is a false positive example;
[0071] Recall refers to the proportion of samples that are correctly predicted as positive among all samples that are actually positive. The calculation formula is:
[0072]
[0073] Where FN is a false negative example;
[0074] The F1 score is the harmonic mean of precision and recall, and its calculation formula is:
[0075]
[0076] Cross-validation divides the data set into K subsets, and uses K-1 subsets as training sets each time, and the remaining one as a test set. This is repeated K times, and the average is taken as the evaluation result of the model. Hyperparameter tuning When there are multiple hyperparameters in the model, such as C (regularization parameter), kernel function type (kernel), and kernel function parameters (such as gamma), the optimal parameter setting can be found by traversing the given hyperparameter combination. Through the above refinement steps, the linear support vector machine model can be comprehensively evaluated, thereby selecting the model configuration that best suits a specific problem.
[0077] Finally, the accuracy, efficiency and practicality of construction project risk assessment can be significantly improved through optimized scheduling methods, so as to better serve the safety management of engineering projects. The optimized scheduling methods include optimized scheduling of risk identification and classification: regularly updating the risk library, incorporating new risk factors and changes, ensuring the timeliness and completeness of risk identification, and using machine learning technology to automatically classify risks to improve the accuracy and efficiency of classification;
[0078] Optimized scheduling of data collection and preprocessing: Implement a real-time data collection system to ensure the timeliness and accuracy of data; establish a data quality inspection mechanism to automatically identify and correct data errors to ensure the quality of data preprocessing;
[0079] Optimized scheduling of model training and validation: Develop automated processes to select the most appropriate model based on data characteristics, use grid search, Bayesian optimization and other methods to automatically adjust model hyperparameters to improve model performance, and reasonably arrange the cross-validation process to ensure the comprehensiveness and efficiency of model validation;
[0080] Optimized scheduling of risk assessment and result analysis: Batch risk assessment of multiple engineering projects to improve assessment efficiency;
[0081] Optimized scheduling of risk response and early warning: Establish an automatic early warning system based on thresholds, and issue an immediate warning once the risk assessment results exceed the predetermined threshold;
[0082] Optimize and schedule the overall process: automate the entire risk assessment process, reduce manual intervention, improve overall efficiency, dynamically adjust resource allocation according to project requirements and risk assessment results, ensure that key resources are used for high-risk projects, establish a feedback loop, compare and analyze the results of risk assessment with actual safety accidents, and continuously improve risk assessment models and methods.
[0083] The above-mentioned specific embodiments are only several preferred embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above-mentioned embodiments, those skilled in the art can make various alternative improvements and combinations to the above-mentioned specific embodiments.
[0084] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. A construction project risk assessment method based on artificial intelligence, comprising a data collection and preprocessing module, a feature selection module, a model building and training module and a model evaluation module, characterized in that: The data collection and preprocessing module processes the data related to the construction project, the feature selection module selects relevant feature data from the processed data set and inputs it into the model construction and training module for training and optimizing model parameters, and the performance of the model is evaluated through the model evaluation module using accuracy, recall rate, and F1 score indicators.
2. The method for risk assessment of construction projects based on artificial intelligence according to claim 1, characterized in that: The data collection and preprocessing module includes a data collection module and a data preprocessing module. The data collection module includes determining data requirements, designing a data collection plan, implementing data collection, data storage and management, data quality checking and data integration. Through these detailed data collection steps, it can be ensured that high-quality data is collected, providing a solid foundation for construction project risk assessment.
3. The method for risk assessment of construction projects based on artificial intelligence according to claim 2 is characterized in that: The data preprocessing module includes data cleaning, denoising and normalization. The data cleaning uses the deduplication function in the database or programming language to ensure that all duplicate data rows are deleted and unique data are retained. The denoising uses filters or smoothing techniques (such as moving average, median filtering) to reduce random noise in the data. For time series data, the difference method can be used to eliminate trend and seasonal effects. The normalization converts the data into a distribution with zero mean and unit variance, scales the data to a specific range, and scales different features to ensure that they have the same importance in model training. Through these refined steps, the quality of the data can be effectively improved, laying a solid foundation for subsequent data analysis and modeling.
4. The method for risk assessment of construction projects based on artificial intelligence according to claim 1, characterized in that: The feature selection module includes feature recognition, feature extraction, feature analysis, feature selection algorithm and feature selection result verification. Through these steps, the feature selection process can be refined to ensure that the final selected feature set can effectively predict the risks of construction projects.
5. The method for risk assessment of construction projects based on artificial intelligence according to claim 4 is characterized in that: The feature selection algorithm adopts filtering selection, which selects features based on univariate statistical tests. The univariate statistical test adopts ANOVA, i.e., analysis of variance. The ANOVA includes the following hypotheses: Null hypothesis (H0): There is no significant difference in the means between the groups; Alternative hypothesis (Ha): There is a significant difference between the means of at least one group and the means of the other groups; Collect data: determine the experimental design, collect data from each group, and ensure that the data meet the assumptions of ANOVA, such as normality and homogeneity of variance; Calculate the total sum of squares (SST): The total sum of squares is the sum of the squares of the differences between all observations and their grand mean, which represents the total variation of the data; Calculate the between-group sum of squares (SSB): The between-group sum of squares is the sum of the squares of the differences between the means of each group and its grand mean, multiplied by the sample size of each group, which represents the variation caused by the differences in means between different groups; Calculate the within-group sum of squares (SSW): The within-group sum of squares is the sum of the squares of the differences between the observations within each group and their group means, which represents the variation within each group; Calculate the mean square (MS): The between-group mean square (MSB) is the between-group sum of squares divided by the between-group degrees of freedom. The within-group mean square (MSW) is the within-group sum of squares divided by the within-group degrees of freedom; Calculate the F value: The F value is the ratio of the mean square between groups to the mean square within groups, that is, F = MSB / MSW; Determine the P value: Use an F distribution table or statistical software to determine the P value of the F value. The P value represents the probability of observing the current F value or a more extreme situation when the null hypothesis is true. Make a decision: If the P value is less than the significance level (usually 0.05), the null hypothesis is rejected, and it is believed that the mean of at least one group is significantly different from the means of other groups. If the P value is greater than the significance level, the null hypothesis cannot be rejected, and it is believed that there is no significant difference in the means between the groups.
6. The method for risk assessment of construction projects based on artificial intelligence according to claim 4 is characterized in that: The model building and training module includes model building and model training. The model building adopts a linear support vector machine model, which performs well when processing high-dimensional data and can effectively classify risks. The steps of model training include function interval, geometric interval and interval maximization.
7. The method for risk assessment of construction projects based on artificial intelligence according to claim 1, characterized in that: The model evaluation module includes evaluation indicators, cross-validation and hyperparameter tuning. The evaluation indicators include accuracy, precision, recall and F1 score. Through the above-mentioned refinement steps, the linear support vector machine model can be comprehensively evaluated to select the model configuration that best suits the specific problem.