A Smart Construction Site Safety Risk Assessment Method Based on Data Mining

Abstract

The present invention discloses a method for intelligent construction site safety risk assessment based on data mining, which relates to the technical field of data mining. The present invention includes the following steps: 1. Obtain various safety-related data sets of real construction sites; 2. On the basis of fully studying the safety risk assessment process and elements of construction sites, construct a safety evaluation system for construction sites; 3. Conduct data analysis based on the data sets of various indicators in the constructed safety evaluation system for construction sites; 4. Improve the AHP method and the entropy weight method based on data by using the D-S synthesis algorithm based on weight assignment and matrix analysis to calculate the subjective and objective weights of each indicator in the indicator layer of the evaluation system; 5. Use the improved D-S evidence fusion algorithm to synthesize multi-source evidence to obtain the indicator weights; 6. Calculate the comprehensive evaluation index of the construction site. The present invention can effectively evaluate the safety of construction site construction, reduce the uncertainty of evaluation results, and improve the credibility of risk assessment results.

Description

Technical Field

[0001] The present invention belongs to the technical field of risk assessment, and more specifically, relates to a method for intelligent construction site safety risk assessment based on data mining. Background Art

[0002] An intelligent construction site is a technology for implementing intelligent management of the construction site of a building project. As a new type of project management method, it aims to optimize the construction process, improve work efficiency and safety, reduce costs and environmental impacts through means such as data analysis, real-time monitoring, and prediction.

[0003] Specifically, intelligent construction site management obtains construction site data through technical means such as sensors and monitoring cameras, and performs data mining and analysis on this data in a virtual reality environment with the project information collected by the Internet of Things, provides process trend prediction and expert preplans, realizes visual intelligent management of project construction, helps enterprises optimize management, improve production efficiency and reduce costs, and thus gradually realizes green construction and safe construction.

[0004] However, at present, there are deficiencies in effectively mining and analyzing the data obtained by the intelligent construction site platform. Existing methods have problems such as one-sided weight assignment and lack of comprehensive consideration of safety situation factors. Therefore, it is of great practical significance and commercial value to develop an intelligent construction site data analysis algorithm to uniformly manage and analyze the data of each intelligent construction site and realize more efficient, safer, and more environmentally friendly building construction.

[0005] After retrieval, the Chinese patent application number is 2022109159361, and the application publication date is September 6, 2022. The invention creation name is: A method and system for safety risk assessment of a construction site; this application first constructs an index system for safety risk assessment of a construction site, the index system includes a number of first-level indicators, and each of the first-level indicators corresponds to a number of second-level indicators respectively; scores are given to the first-level indicators and the second-level indicators respectively, and the membership degree value of each indicator to each construction safety level is calculated according to the scoring result of each indicator, and a membership degree matrix is constructed based on the membership degree values of each indicator; the objective weight of each indicator is determined according to the entropy weight method, and at the same time, the subjective weight of each indicator is determined according to the analytic hierarchy process, and the comprehensive weight of the indicator is calculated according to the objective weight and subjective weight of each indicator; the construction site safety score is calculated according to the membership degree matrix and the comprehensive weight; according to the safety score, combined with the construction safety level division table, the construction safety level of the current construction site is determined. Although this application assigns weights to the second-level indicators through subjective and objective methods, it does not consider the impact of the differences in the experience values of experts on the evaluation results, and the processing of assigning subjective and objective weights in this application is just a simple linear combination, and the connection between the subjective and objective weight coefficients and the real data is not clear. Summary of the Invention

[0006] 1. Technical Problem to be Solved by the Invention

[0007] In view of the deficiencies in the existing technology, the present invention provides a method for evaluating the safety risks of a smart construction site based on data mining. The present invention uses the improved D-S evidence fusion theory to assign weights to safety indicators, thereby reducing the uncertainty in the process of assigning weights to safety indicators, and can more effectively evaluate the safety situation at the construction site, so as to facilitate managers to more directly grasp the on-site safety issues.

[0008] 2. Technical Solution

[0009] To achieve the above object, the technical solution provided by the present invention is as follows:

[0010] A method for evaluating the safety risks of a smart construction site based on data mining according to the present invention comprises the following steps:

[0011] Step 1: Obtain data through sensors, monitoring cameras, and GPS positioning, and determine the indicators affecting the safety of the construction site;

[0012] Step 2: Based on a full study of the safety risk assessment process and elements of the construction site, establish a safety evaluation system for the construction site based on the indicators of the construction site safety determined in Step 1;

[0013] Step 3: According to the system established in Step 2 and the questionnaire survey of industry experts, improve the analytic hierarchy process using the D-S synthesis algorithm based on weight assignment and matrix analysis to determine the subjective weight coefficients of the indicators in the safety evaluation system of the construction site;

[0014] Step 4: According to the data set obtained in Step 1 and the system established in Step 2, use the entropy weight method to determine the objective weight coefficients of the indicators in the safety evaluation system of the construction site;

[0015] Step 5: According to the output results of Step 3 and Step 4, use the improved D-S evidence fusion theory to determine the combined weight coefficients;

[0016] Step 6: According to the data set in Step 1 and the combined weight coefficients determined in Step 5, calculate the safety of the construction site of the evaluation object according to the TOPSIS evaluation algorithm.

[0017] 3. Beneficial Effects

[0018] Adopting the technical solution provided by the present invention, compared with the existing well-known technologies, has the following remarkable effects:

[0019] (1) A method for safety risk assessment of intelligent construction sites based on data mining according to the present invention aims at the problem that it is difficult to assign weights to risk factors at the construction site during the safety risk assessment process of construction sites. To evaluate the overall situation of the construction site, the D-S evidence synthesis algorithm that combines weight assignment and matrix analysis is used to fuse the differences in safety risks among multiple experts, and the subjective weights of safety risk indicators obtained are more scientific and reasonable, avoiding the influence of the differences in experts' experience values on the evaluation results;

[0020] (2) A method for safety risk assessment of intelligent construction sites based on data mining according to the present invention uses the entropy weight method to calculate the objective weights of safety risk indicators for sample data, and through an improved D-S evidence theory algorithm, the combined weight assignment method of the AHP method and the entropy weight method is used. Starting from the subjective and objective perspectives respectively, it overcomes the one-sidedness of the evaluation results brought by a single weight assignment method, and the combined method not only reduces the subjectivity of the expert experience weight assignment of the AHP method, but also reduces the fluctuation of the weight value caused by data changes; the improved D-S evidence theory can better fuse the subjective and objective weights.

[0021] (3) Since different evidences show different authorities or reliabilities, when applying the traditional D-S evidence theory to make decisions, it is more reasonable to handle low-conflict evidences, but it has the drawback of veto power when encountering high conflicts. For example, the Zadeh paradox problem. The weight fusion method proposed by the present invention takes into account the credibility of evidences, and by introducing a method for solving the discounted weight, associates the connections between evidence sources, thereby obtaining the discounted weight and adjusting the weight, and linearly combines them to obtain a comprehensive weight coefficient, and obtains the discounted credibility function obtained by each evidence under each proposition subset; this method determines the weight coefficient considering the trust degree between evidences, avoiding the artificial definition of evidence weight coefficients.

[0022] (4) A method for safety risk assessment of intelligent construction sites based on data mining according to the present invention can effectively handle highly conflicting evidences. In the case of fewer evidence sources, it can converge to the correct result, reducing the decision-making risk and improving the credibility of the risk assessment results. From a new perspective, objective data and empirical knowledge are more realistically integrated into the comprehensive evaluation of construction site safety risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a schematic diagram of the data mining and analysis process of the intelligent construction site platform of the present invention;

[0024] Figure 2 is a schematic diagram of the construction site safety evaluation system constructed by the present invention;

[0025] Figure 3 is a schematic diagram of the comparison of the combined weight coefficient results of the present invention. Detailed implementation mode

[0026] To further understand the content of the present invention, the present invention will be described in detail in combination with the accompanying drawings and embodiments.

[0027] Embodiment 1

[0028] Combined with Figure 1 , a method for evaluating safety risks of an intelligent construction site based on data mining in this embodiment specifically includes the following steps:

[0029] Step 1: Obtain data through sensors, monitoring cameras, and GPS positioning, respectively as the input of the objective weight of the acquisition index and the TOPSIS algorithm, and determine the indicators affecting the safety of the construction site.

[0030] Step 2: Based on a full study of the safety risk assessment process and elements of the construction site, establish a safety evaluation system for the construction site based on the indicators of the safety of the construction site determined in Step 1.

[0031] Refer to Figure 2 , the safety evaluation system for the construction site in this embodiment includes an objective layer, a criterion layer (primary index), and an index layer (secondary index). Among them, the objective layer is the risk U of the construction site construction, and the criterion layer is the personnel management index U1, the project safety management index U2, the supervision safety inspection index U3, the mechanical equipment safety index U4, and the safety index U5 of dangerous and major projects. The index layer is education and training U11, reward and punishment behavior U12, AI video detection of human behavior U13, opening order number U21, mechanical equipment monitoring number U22, construction safety management U31, civilized construction inspection U32, scaffolding inspection U33, high-altitude operation inspection U34, construction electricity inspection U35, formwork engineering inspection U36, crane inspection U37, tower crane warning U41, lift warning U42, unloading platform warning U43, deep foundation pit warning U51, high formwork warning U52, and edge warning U53.

[0032] Step 3: According to the system established in Step 2 and the questionnaire survey of industry experts, use the D-S synthesis algorithm based on weight assignment and matrix analysis to improve the analytic hierarchy process, and determine the subjective weight of the indicators in the safety evaluation system of the construction site;

[0033] The implementation steps of the improved analytic hierarchy process are as follows:

[0034] 3-1: In the form of an expert questionnaire, for the five indicators in the criterion layer, use the 3-scale method, and score by 3 experts. Compare the safety levels among the primary indicators in the criterion layer. According to the scoring results of the three experts, respectively construct the judgment matrices A 1 、A 2 、A 3 of the three experts for the primary indicators in the criterion layer, specifically as follows:

[0035]

[0036] 3-2. Calculate the three judgment matrices respectively to obtain the maximum eigenvalue λ of each judgment matrix max and the eigenvector β, and conduct a consistency test. The consistency ratio CR of matrix A 1 is 0.0023, the consistency ratio CR of matrix A 2 is 0.0029, and the consistency ratio CR of matrix A 3 is 0.0206. All consistency ratios are less than 0.1, meeting the consistency requirements of the judgment matrix. The subjective index weight w an , where n = 1, 2, 3, is the result of normalizing the eigenvector β. The calculation results are shown in Table 1:

[0037]

[0038] 3-3. Adopt the D-S evidence synthesis algorithm based on weight assignment and matrix analysis to fuse the subjective weights of empirical values determined by different experts, and the subjective weights w' of each first-level index in the criterion layer are obtained through evidence fusion a . Among them, obtained from step 3-2, the subjective weights obtained by three experts for the first-level indicators of the criterion layer are w a1 , w a2 , w a3 . Let matrices A, B, and C represent the subjective weights w a1 , w a2 , w a3 , respectively, that is:

[0039] A = w a1 = (0.2701, 0.1223, 0.1619, 0.2228, 0.2228),

[0040] B = w a2 = (0.3151, 0.1088, 0.1487, 0.2142, 0.2132),

[0041] C = w a3 = (0.2992, 0.1180, 0.1512, 0.2293, 0.2022).

[0042] Matrix M1 = A T ×B. The sum of the non-diagonal elements in M1 is the conflict degree K between expert A 1 and expert A 2 .

[0043] Multiply the diagonal elements of M1 with matrix C in the same way to obtain M2. That is, the sum of the non-diagonal elements in M2 is the conflict degree K of expert A 1 、A 2 、A 3 。

[0044] The improved D-S evidence combination formula using weight distribution:

[0045]

[0046] where f(θ) = K·q(θ), indicating the average support degree of all evidence sources for θ.

[0047] Due to the difference in the experience values of experts, through calculation, the experience values of experts are fused to obtain the combined subjective weight w a =(0.3051, 0.1120, 0.1496, 0.2216, 0.2114).

[0048] Similarly, obtain the weights of the secondary indicators of each index layer relative to the primary indicators of the construction site safety risk evaluation criterion layer. Multiply the subjective weights of each index in the criterion layer with the subjective weights of the corresponding indicators in the index layer. Finally, obtain the comprehensive subjective weight w' of all secondary indicators relative to the target layer a =(0.1403, 0.0973, 0.0675, 0.0746, 0.0373, 0.0536, 0.0262, 0.0107, 0.0124, 0.0049, 0.0049, 0.0369, 0.0865, 0.0760, 0.0591, 0.0851, 0.0729, 0.0535).

[0049] Step 4: Based on the dataset obtained in Step 1 and the system established in Step 2, use the entropy weight method to determine the objective weight coefficient w' of the indicators in the construction site safety evaluation system e ;

[0050] Through the on-site index data of each construction site obtained from the intelligent construction site platform in Step 1, select 5 construction site evaluation objects and the data of 18 secondary indicators to construct the decision matrix X = [x ij 5×18 ,and standardize the data of X. Finally, obtain the standardized matrix R:

[0051]

[0052] From the standardized matrix R, using the entropy weight method, the entropy values e of 18 indicators can be calculated j ​=(0.9963, 0.9856, 0.9878, 0.9986, 0.9981, 0.9229, 0.9942, 0.9505, 0.9327, 0.9722, 0.6114, 0.8199, 0.9593, 0.9420, 0.9317, 0.9621, 0.9764, 0.9803), from which the objective weight w’ e =(0.0034, 0.0134, 0.0113, 0.0013, 0.0017, 0.0714, 0.0054, 0.0459, 0.0624, 0.0258, 0.3602, 0.1670, 0.0378, 0.0545, 0.0633, 0.0351, 0.0219, 0.0183).

[0053] Step 5: According to the output results of Step 3 and Step 4, use the improved D-S evidence fusion theory to determine the combined weight coefficient;

[0054] 5-1: Take the subjective weight w' a obtained in Step 3 and the objective weight w' e obtained in Step 4 as two different evidence sources, that is, the subjective weight w' a and the objective weight w' e as two different evidence sources E = {e 1 , e 2};

[0055] Construct the recognition frame Θ = {θ 1 , θ 2 , …, θ 18}, where θ n represents the set composed of each secondary index in the index layer, and the recognition frame satisfies Among them, the basic probability function m represents the degree of support of all evidence sources for the subset θ n .

[0056] Use the distance function to calculate the discount weight between the subjective weight w’ a and the objective weight w’ e ;

[0057] Use to calculate the discount weight configuration coefficient; in this embodiment, the discount weight configuration coefficients are w 1 = 0.6533 and w 2 = 0.3477.

[0058] Use to calculate the weighted average evidence similarity of all evidence, which is the adjusted weight; in this embodiment, the adjusted weight s = [0.5110, 0.4890]T 。

[0059] Use α i = λw i +(1 - λ)s to calculate the comprehensive weight coefficient of each evidence source; in this embodiment, λ = 0.5, and the comprehensive weight α = [0.5831, 0.4169] T 。

[0060] Use to calculate the discounted belief function obtained by each evidence ei under each proposition subset θ n ;

[0061] Use to calculate the probability assignment function of each proposition subset θ i after fusing the data of each evidence source, that is, the combined weight coefficient; in this embodiment, the combined weight coefficient w j = (0.0832, 0.0623, 0.0441, 0.0440, 0.0225, 0.0610, 0.01753, 0.0254, 0.0332, 0.0136, 0.1530, 0.0911, 0.0662, 0.0670, 0.0609, 0.0643, 0.05164, 0.03883).

[0062] In the above formula, D is the difference between w 1 and w 2 , that is, the difference between the weight configuration coefficients, and β ij is the initial belief, that is, the weight under each evidence source, is the probability assignment function.

[0063] Step 6: According to the dataset obtained in Step 1 and the combined weight coefficient determined in Step 5, calculate the construction site safety of the evaluation object according to the TOPSIS evaluation algorithm.

[0064] From the on-site index data of each construction site obtained from the intelligent construction site platform in Step 1, select 5 evaluation objects and 18 index data to construct a decision matrix X = [x ij 5×18

[0065] Construct a weighted decision matrix from the combined weights of each index in the index layer calculated in Step 5; the formula for the weighted decision matrix is as follows:

[0066]

[0067] Use to calculate the Euclidean space distances of each evaluation object to the "positive ideal solution" and the "negative ideal solution";

[0068] Use​ Calculating the relative proximity is the situation of advantages and disadvantages.

[0069] In the above formula, x’ ij represents the normalized data of the jth safety index on the ith object, and w j is the combined weight of the jth index, and Z + represents the positive ideal solution of the weighted decision matrix, which is composed of the maximum values of each index parameter on all evaluation objects; Z - represents the negative ideal solution of the weighted decision matrix, which is composed of the minimum values of each index parameter on all evaluation objects.

[0070] The closer the relative proximity is to 1, the closer it is to the optimal level, and the higher the safety of the construction site. The calculation results are shown in Table 2:

[0071]

[0072] The values S of the relative proximity i are respectively (0.5133, 0.7131, 0.4484, 0.4825, 0.4783), and the safety ranking of the 5 evaluation items is b > a > d > e > c: among them, project b has the highest score, indicating that the risk of the construction site in this area is the smallest, which is consistent with the actual situation, reflecting that the method proposed in the present invention can better obtain a credible comprehensive evaluation result when the evidence information is clear and reliable.

[0073] The above schematically describes the present invention and its implementation manners. This description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. Therefore, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments without creative efforts without departing from the gist of the present invention, they shall fall within the protection scope of the present invention.

Claims

1. A method for safety risk assessment of intelligent construction sites based on data mining, characterized in that, its steps are as follows: Step 1: Obtain data through sensors, monitoring cameras, and GPS positioning, and determine the indicators affecting the safety of construction sites; Step 2: Based on a full study of the safety risk assessment process and elements of construction sites, establish a safety evaluation system for construction sites based on the indicators of construction site safety determined in Step 1; Step 3: According to the system established in Step 2 and the questionnaire survey of industry experts, improve the analytic hierarchy process using the D-S synthesis algorithm based on weight assignment and matrix analysis to determine the subjective weight coefficients of the indicators in the safety evaluation system of construction sites; the specific steps are as follows: 3-1: For the five indicators in the criterion layer, according to the expert scoring results, use the 3-scale method to compare the safety levels between indicators and construct the judgment matrix of each indicator in the criterion layer; 3-2. Calculate and solve the maximum eigenvalue λ of the judgment matrix max and the eigenvector β, and conduct a consistency test. Calculate the consistency ratio CR. When the consistency ratio CR is less than 0.1, it indicates that the consistency test passes; otherwise, the judgment matrix needs to be reconstructed. Among them, the obtained eigenvector β is normalized to obtain the index weight w a ; 3-3. Adopt the D-S evidence synthesis algorithm based on weight assignment and matrix analysis to fuse the subjective weights of empirical values determined by different experts. Through evidence fusion, obtain the subjective weights of each index in the criterion layer and the subjective weights of the index layer relative to the first-level index in the criterion layer. Multiply the weights obtained from the criterion layer and the index layer to get the comprehensive subjective weight w' of all secondary indexes relative to the target layer a ; Step 4: Based on the dataset obtained in Step 1 and the system established in Step 2, use the entropy weight method to determine the objective weight coefficients of the indicators in the safety evaluation system of construction sites; Step 5: According to the output results of Step 3 and Step 4, use the improved D-S evidence fusion theory to determine the combined weight coefficients; the specific process is as follows: Take the subjective and objective weights obtained in Step 3 and Step 4 as the inputs of the combined weights of different evidence sources in Step 5, that is, the subjective weight w' a and the objective weight w' e as two different evidence sources E = {e 1 , e 2}; Construct an identification framework Θ = {θ 1 , θ 2 , …, θ 18}, where θ n represents the set composed of each secondary index in the index layer, and the identification framework satisfies Among them, the basic probability function m represents the degree of support of all evidence sources for the subset θ n . Use the formula: where D is the difference between w 1 and w 2 , that is, the difference between weight configuration coefficients; Obtain the subjective weight w' a and the objective weight w' e to obtain the discount weight configuration coefficient w 1 between w 2 ; Substitute the calculated discounted weight coefficients into the formula: Calculate the weighted average evidence similarity and the comprehensive weight of all evidence. Finally, after fusing the data of each evidence source, substitute it into the formula to obtain the combined weight coefficient; Step 6: Based on the dataset in Step 1 and the combined weight coefficients determined in Step 5, calculate the safety of the construction site of the evaluation object according to the TOPSIS evaluation algorithm.

2. A method for safety risk assessment of intelligent construction sites based on data mining according to claim 1, characterized in that: The safety evaluation system for construction sites established in Step 2 includes an objective layer, a criterion layer, and an indicator layer. The objective layer is the risk U of the construction site construction; the criterion layer includes personnel management indicators U1, project safety management indicators U2, supervision safety inspection indicators U3, mechanical equipment safety indicators U4, and major dangerous project safety indicators U5; the indicator layer includes education and training U11, reward and punishment behavior U12, AI video detection of human behavior U13, number of work orders issued U21, number of mechanical equipment monitored U22, construction safety management U31, civilized construction inspection U32, scaffolding inspection U33, high-altitude operation inspection U34, construction electricity inspection U35, formwork project inspection U36, crane inspection U37, tower crane warning U41, lift warning U42, unloading platform warning U43, deep foundation pit warning U51, high formwork warning U52, and edge warning U53.

3. A method for safety risk assessment of intelligent construction sites based on data mining according to claim 2, characterized in that: In step 3-3, based on the index weight w obtained in step 3-2 a , the weight conflict degree K between different experts is obtained, and the improved D-S evidence synthesis formula after weight distribution is used: where f(θ) = K·q(θ), represents the average degree of support for θ by all evidence sources; Integrate the expert experience values to obtain the combined subjective weights of each index in the criterion layer; similarly, calculate the subjective weights of the secondary indexes in each index layer with respect to the primary indexes in the construction site safety risk evaluation criterion layer; multiply the subjective weights of each index in the criterion layer by the subjective weights of the corresponding indexes in each index layer. Finally, obtain the comprehensive subjective weight w' of all secondary indexes with respect to the target layer a .

4. A method for safety risk assessment of intelligent construction sites based on data mining according to claim 3, characterized in that: In step 4, first select the on-site index data of each construction site obtained from the intelligent construction site platform to construct an evaluation index matrix, and standardize it to obtain a standardized index matrix R; from the standardized matrix R, calculate the entropy value e of each index j and the objective weight value w' e .

5. A method for safety risk assessment of intelligent construction sites based on data mining according to claim 4, characterized in that: The specific process of Step 6 is as follows: (1) Construct a decision matrix from the on-site indicator data of each construction site obtained from the intelligent construction site platform in Step 1. (2) Adopt Calculate the weighted decision matrix; (3) Adopt Calculate the distances from each evaluation object to the positive and negative ideal solutions; (4) Adopt Calculate the relative fitting degree to determine the quality situation; In the above formula, x′ ij represents the normalized data of the j-th safety index on the i-th object, and w j is the combined weight of the j-th index. Z + represents the positive ideal solution of the weighted decision matrix, which is composed of the maximum values of each index parameter on all evaluation objects; Z - represents the negative ideal solution of the weighted decision matrix, which is composed of the minimum values of each index parameter on all evaluation objects.

Images