Offshore bridge construction safety risk identification method based on text mining and N-K-SNA model

Abstract

The invention discloses an offshore bridge construction safety risk identification method based on text mining and an N-K-SNA model, and the method comprises the steps: 1) obtaining an offshore bridge construction accident report, and carrying out the preprocessing of the accident report, and obtaining an accident report text; 2) respectively adding the accident report text into a special dictionary for offshore bridge construction and a stop word dictionary, and then carrying out word segmentation to obtain the accident report text after word segmentation; 3) carrying out keyword information extraction on the report text subjected to word segmentation by utilizing a TF-IDF (Term Frequency-Inverse Document Frequency) algorithm; 4) combining the keywords into risk factors, and dividing the risk factors into four categories of personnel, articles, environment and management factors; 5) carrying out statistics on risk coupling types and occurrence frequencies of the collection cases, and calculating a coupling value of a target risk coupling mode by using an N-K model; 6) constructing a risk factor mutual occurrence matrix, generating a risk factor network, and analyzing the centrality and accessibility of network nodes through an SNA model; and 7) fusing the N-K model and the SNA model, and identifying key risk factors of offshore bridge construction.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore bridge construction safety risk identification, and in particular to an offshore bridge construction safety risk identification method based on text mining and NK-SNA model. Background Art

[0002] As an important part of transportation infrastructure, offshore bridges have played a vital role in promoting the rapid development of my country's economy. Due to the existence of various uncertain risk factors in the construction of offshore bridges, the risk factors in offshore bridge construction are not only numerous, but also mutually coupled, resulting in great safety hazards in the construction of offshore bridges. Therefore, it is very necessary to carry out construction safety management of offshore bridges.

[0003] In recent years, domestic and foreign scholars have conducted a lot of research on risk management in the construction process of offshore bridges. At present, common risk identification methods include expert investigation, WBS-RBS, literature investigation, and fault tree analysis. However, traditional risk factor identification methods rely heavily on the subjective dependence of experts. With the increasing complexity of the modern industrial system, the accuracy of professional risk identification cannot be well guaranteed, which leads to certain limitations in the results of risk factor identification. At the same time, the occurrence of offshore bridge construction accidents is often not caused by a single disaster factor, but the result of the coupling of multiple risk disaster factors. The current offshore bridge construction risk coupling assessment model is difficult to evaluate both the top risk factors and the bottom risk factors at the same time. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method for identifying safety risks in offshore bridge construction based on text mining and NK-SNA model, which can accurately and objectively identify key safety risk factors in offshore bridge construction.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0006] A method for identifying safety risks in offshore bridge construction based on text mining and NK-SNA model, including:

[0007] Step 1): Obtain an accident report on offshore bridge construction, and pre-process the accident report to obtain a pre-processed accident report text;

[0008] Step 2): adding the pre-processed accident report text to the offshore bridge construction special dictionary and the stop word dictionary respectively, and then performing word segmentation to obtain the accident report text after word segmentation;

[0009] Step 3): Use the TF-IDF algorithm to extract keyword information from the report text after word segmentation;

[0010] Step 4): Combine keywords into risk factors through professional knowledge of offshore bridge construction industry, and divide risk factors into four categories: personnel factors, material factors, environmental factors, and management factors;

[0011] Step 5): Statistically collect the risk coupling types and occurrence frequencies of the cases, and use the NK model to calculate the coupling value of the target risk coupling mode;

[0012] Step 6): Construct the mutual occurrence matrix of risk factors, generate the risk factor network, and analyze the centrality and accessibility of network nodes through the SNA model;

[0013] Step 7): Integrate the NK model and the SNA model to identify the key risk factors of offshore bridge construction.

[0014] As a possible implementation, further, step 1) specifically includes the following steps:

[0015] Use Python to request and parse tags of offshore bridge engineering accident web pages on the Internet, analyze and collect relevant accident case data;

[0016] The accident case data is uniformly organized into txt format files to obtain the pre-processed accident case report text.

[0017] As a possible implementation, further, step 2) specifically includes the following steps:

[0018] Through the word segmentation library in Python and the constructed offshore bridge engineering text dictionary, the pre-processed accident report text is stopped and segmented, punctuation marks and meaningless words are removed, and the processed report is output as a txt format file.

[0019] As a possible implementation, further, step 3) specifically includes the following steps:

[0020] Step 3.1): Calculate the TF-IDF value of each word in the accident report text;

[0021] Step 3.2): Introduce the inverse document frequency (IDF) and use the TF-IDF algorithm to identify words that appear less frequently in the article but contribute significantly to the text;

[0022] Step 3.3): Delete words with large TF-IDF values ​​that are not helpful for identifying risk factors of offshore bridge construction, and finally select the first 100 processed words and sort them according to TF-IDF values.

[0023] As a possible implementation, further, step 3.1) is specifically as follows:

[0024] First, we conduct word frequency statistics to count the number of times each word appears in the accident report;

[0025] Then the frequency of these words is divided by the total number of times in the accident report text to obtain the standardized word frequency TF, where the expression of TF is as follows:

[0026]

[0027] In the above formula, f t,d Represents the number of times word t appears in document d; ∑ t'∈d f t',d Indicates the total words of the accident report text.

[0028] As a possible implementation, further, the expression of IDF in step 3.2) is as follows:

[0029]

[0030] In the above formula, N represents the total number of documents in the document collection; |{d∈D:t∈d}| represents the number of documents containing term t.

[0031] As a possible implementation, further, step 4) specifically includes the following steps:

[0032] Combine the obtained risk factor keywords with professional knowledge for contextual processing to identify security risk factors;

[0033] Words with similar meanings were merged, and finally the risks were divided into four categories: personnel factors, object factors, environmental factors, and management factors.

[0034] As a possible implementation, further, step 5) specifically includes the following steps:

[0035] The collected accident cases were classified into risk coupling forms, including single-factor coupling, double-factor coupling and multi-factor coupling, and their occurrence frequencies were counted;

[0036] The NK model was used to calculate the risk coupling values ​​of the two-factor coupling mode and the multi-factor coupling mode.

[0037] As a possible implementation, further, step 6) specifically includes the following steps:

[0038] Step 6.1): Check the accident report text and obtain the cause of the accident in each text;

[0039] Step 6.2): ​​Analyze the relative relationship between various influencing factors in the accident report text;

[0040] Step 6.3): Mark risk factors that are related or likely to occur together as co-occurring risk factors;

[0041] Step 6.4): Generate a risk factor network using the risk factor mutual occurrence matrix, and perform betweenness centrality analysis on the generated network;

[0042] The expression of betweenness centrality is as follows:

[0043]

[0044] In the above formula, C b (i) represents the betweenness centrality of the node, σ st is the number of shortest paths from node s to node t; σ st (i) is the number of shortest paths from node s to node t that pass through node i.

[0045] As a possible implementation, further, step 7) specifically includes the following steps:

[0046] The risk factor coupling value obtained in step 5) is used as the correction coefficient to correct the betweenness centrality value obtained in step 6) to obtain the corrected betweenness centrality value. The correction formula is as follows:

[0047] C b '(i)=T i ×C b (i) (7)

[0048] In the above formula, C b '(i) represents the corrected betweenness centrality; T i Indicates the risk coupling value of the coupling mode corresponding to the risk factor; C b (i) represents the betweenness centrality of the node;

[0049] By comparing the betweenness centrality before and after correction, the risk factors whose values ​​are amplified are key risk factors, and the risk factors whose values ​​are reduced are secondary risk factors.

[0050] By adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0051] The present invention provides a method for identifying risk factors of offshore bridge construction accidents based on text mining and NK-SNA, which can provide a more comprehensive, objective and in-depth understanding and mining of risk factors in texts. By combining the NK model with the SNA model, the influence of key risk factors is amplified, the influence of secondary risk factors is reduced, and the key factors are made more prominent, thereby effectively improving the accuracy and effectiveness of risk identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 It is a simple flow chart of the present invention. DETAILED DESCRIPTION

[0054] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0055] See attached Figure 1 As shown, this embodiment provides a method for identifying safety risks in offshore bridge construction based on text mining and NK-SNA model, including:

[0056] Step 1): Obtain an accident report on offshore bridge construction, and pre-process the accident report to obtain a pre-processed accident report text, specifically including the following steps:

[0057] Python is used to request and parse the labels of offshore bridge engineering accident web pages contained in domestic and foreign media news and local government websites, analyze and collect relevant accident case data, and organize the accident case data into txt format files to obtain the pre-processed accident case report text.

[0058] Step 2): The pre-processed accident report text is added to the offshore bridge construction special dictionary and the stop word dictionary respectively and then segmented to obtain the accident report text after word segmentation, which specifically includes the following steps:

[0059] The pre-processed accident report text is subjected to stopword and word segmentation by using the word segmentation function in Python and the constructed offshore bridge engineering text dictionary, including removing punctuation marks and meaningless words, such as "according to" and "at this time", and the processed report is output as a txt format file.

[0060] Step 3): Extract keyword information from the report text after word segmentation using the TF-IDF algorithm, specifically including the following steps:

[0061] Step 3.1): Calculate the TF-IDF value of each word in the accident report text. The specific steps are as follows: first, perform word frequency statistics to count the number of times each word appears in the accident report;

[0062] Then the frequency of these words is divided by the total number of times in the accident report text to obtain the standardized word frequency TF, where the expression of TF is as follows:

[0063]

[0064] In the above formula, f t,d Indicates the number of times word t appears in document d; Σ t'∈d f t',d Indicates the total words of the accident report text.

[0065] Step 3.2): Introduce the inverse document frequency (IDF) and use the TF-IDF algorithm to identify words that appear less frequently in the article but contribute significantly to the text; the expression of IDF is as follows:

[0066]

[0067] In the above formula, N represents the total number of documents in the document collection; |{d∈D:t∈d}| represents the number of documents containing term t.

[0068] Step 3.3): Delete words with large TF-IDF values ​​that are not helpful for identifying risk factors of offshore bridge construction, such as "bridge erection machine", "scaffolding", "crane", etc. Finally, select the first 100 processed words and sort them according to the TF-IDF value.

[0069] Step 4): Combine keywords into risk factors through professional knowledge of offshore bridge construction industry, and divide risk factors into four categories: personnel factors, material factors, environmental factors, and management factors. The specific steps include the following:

[0070] The obtained risk factor keywords are contextualized with professional knowledge to more completely identify potential safety risk factors; in addition, due to the different sources of different accident cases, the descriptions of the same accident risk factors are also different, so it is necessary to merge these words with similar meanings, such as "inadequate management", "ineffective management", and "insufficiently strict management" can be unified into "inadequate management". Finally, the risks are divided into four categories: personnel factors, material factors, environmental factors, and management factors.

[0071] Step 5): Statistically collect the risk coupling types and frequencies of the cases, and use the NK model to calculate the coupling value of the target risk coupling mode, which specifically includes the following steps:

[0072] The collected accident cases were classified into risk coupling forms, which were divided into single-factor coupling, double-factor coupling and multi-factor coupling, and their occurrence frequencies were counted; the number of occurrences and frequencies of various coupling forms are shown in Table 1 below:

[0073] Table 1 Frequency and frequency of risk coupling of offshore bridge construction accidents

[0074]

[0075] Then, the NK model is used to calculate the risk coupling values ​​of the two-factor coupling method and the multi-factor coupling method. The following provides examples of the two-factor and multi-factor risk coupling formulas for explanation:

[0076] The dual-factor risk coupling expression is as follows:

[0077]

[0078] The three-factor risk coupling expression is as follows:

[0079]

[0080] The four-factor risk coupling expression is as follows:

[0081]

[0082] In the above formula, r, w, h, and g represent personnel, objects, environment, and management factors respectively. rwhg represents the probability of coupling of four risk types: personnel in state r, objects in state w, environment in state h, and management in state g; P rwh represents the probability of coupling of three risk types: personnel in state r, objects in state w, and environment in state h; P rw P represents the probability of coupling of two risk types: personnel in state r and objects in state w; r… , P .w.. , P ..h. , P …g They represent the probability of the human factor being in state r, the probability of the item factor being in state w, the probability of the environment being in state h, and the probability of the management being in state g. In the specific calculation process, the corresponding parameters in formulas (3) to (5) can be replaced to obtain the risk coupling value of each two-factor coupling mode and multi-factor coupling mode.

[0083] Step 6): Construct a risk factor mutual occurrence matrix, generate a risk factor network, and analyze the centrality and accessibility of network nodes through the SNA model, which specifically includes the following steps:

[0084] Step 6.1): Check the accident report text and obtain the cause of the accident in each text;

[0085] Step 6.2): ​​Analyze the relative relationship between various influencing factors in the accident report text;

[0086] Step 6.3): Mark risk factors that are related or likely to occur together as co-occurring risk factors;

[0087] Step 6.4): Generate a risk factor network using the risk factor mutual occurrence matrix, and perform betweenness centrality analysis on the generated network;

[0088] Among them, the constructed mutual occurrence matrix is ​​shown in Table 2 below:

[0089] Table 2 Inter-occurrence matrix of offshore bridge risk factors

[0090]

[0091] The expression of betweenness centrality is as follows:

[0092]

[0093] In the above formula, C b (i) represents the betweenness centrality of the node, σ st is the number of shortest paths from node s to node t; σ st (i) is the number of shortest paths from node s to node t that pass through node i.

[0094] Step 7): Integrate the NK model and the SNA model to identify the key risk factors of offshore bridge construction, which specifically includes the following steps:

[0095] The risk factor coupling value obtained in step 5) is used as the correction coefficient to correct the betweenness centrality value obtained in step 6) to obtain the corrected betweenness centrality value. The correction formula is as follows:

[0096] C b '(i)=T i ×C b (i) (7)

[0097] In the above formula, C b '(i) represents the corrected betweenness centrality; T i Indicates the risk coupling value of the coupling mode corresponding to the risk factor; C b (i) represents the betweenness centrality of the node;

[0098] By comparing the betweenness centrality before and after correction, the risk factors whose values ​​are amplified are key risk factors, and the risk factors whose values ​​are reduced are secondary risk factors.

[0099] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for identifying safety risks in offshore bridge construction based on text mining and NK-SNA model, characterized in that: include: Step 1): Obtain an accident report on offshore bridge construction, and pre-process the accident report to obtain a pre-processed accident report text; Step 2): adding the pre-processed accident report text to the offshore bridge construction special dictionary and the stop word dictionary respectively, and then performing word segmentation to obtain the accident report text after word segmentation; Step 3): Use the TF-IDF algorithm to extract keyword information from the report text after word segmentation; Step 4): Combine keywords into risk factors through professional knowledge of offshore bridge construction industry, and divide risk factors into four categories: personnel factors, material factors, environmental factors, and management factors; Step 5): Statistically collect the risk coupling types and occurrence frequencies of the cases, and use the NK model to calculate the coupling value of the target risk coupling mode; Step 6): Construct the mutual occurrence matrix of risk factors, generate the risk factor network, and analyze the centrality and accessibility of network nodes through the SNA model; Step 7): Integrate the NK model and the SNA model to identify the key risk factors of offshore bridge construction.

2. The method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1 is characterized in that: Step 1) specifically includes the following steps: Use Python to request and parse tags of offshore bridge engineering accident web pages on the Internet, analyze and collect relevant accident case data; The accident case data is uniformly organized into txt format files to obtain the pre-processed accident case report text.

3. The method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1 is characterized in that: Step 2) specifically includes the following steps: Through the word segmentation library in Python and the constructed offshore bridge engineering text dictionary, the pre-processed accident report text is stopped and segmented, punctuation marks and meaningless words are removed, and the processed report is output as a txt format file.

4. The method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1 is characterized in that: Step 3) specifically includes the following steps: Step 3.1): Calculate the TF-IDF value of each word in the accident report text; Step 3.2): Introduce the inverse document frequency (IDF) and use the TF-IDF algorithm to identify words that appear less frequently in the article but contribute significantly to the text; Step 3.3): Delete words with large TF-IDF values ​​that are not helpful for identifying risk factors of offshore bridge construction, and finally select the first 100 processed words and sort them according to TF-IDF values.

5. The method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 4 is characterized in that: Step 3.1) is specifically: First, we conduct word frequency statistics to count the number of times each word appears in the accident report; Then the frequency of these words is divided by the total number of times in the accident report text to obtain the standardized word frequency TF, where the expression of TF is as follows: In the above formula, f t,d Represents the number of times word t appears in document d; ∑ t'∈d f t',d Indicates the total words of the accident report text.

6. The method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 4 is characterized in that: The expression of IDF in step 3.2) is as follows: In the above formula, N represents the total number of documents in the document collection; |{d∈D:t∈d}| represents the number of documents containing term t.

7. According to the method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1, step 4) specifically comprises the following steps: Combine the obtained risk factor keywords with professional knowledge for contextual processing to identify security risk factors; Words with similar meanings were merged, and finally the risks were divided into four categories: personnel factors, object factors, environmental factors, and management factors.

8. According to the method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1, step 5) specifically comprises the following steps: The collected accident cases were classified into risk coupling forms, including single-factor coupling, double-factor coupling and multi-factor coupling, and their occurrence frequencies were counted; The NK model was used to calculate the risk coupling values ​​of the two-factor coupling mode and the multi-factor coupling mode.

9. According to the method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1, step 6) specifically comprises the following steps: Step 6.1): Check the accident report text and obtain the cause of the accident in each text; Step 6.2): ​​Analyze the relative relationship between various influencing factors in the accident report text; Step 6.3): Mark risk factors that are related or likely to occur together as co-occurring risk factors; Step 6.4): Generate a risk factor network using the risk factor mutual occurrence matrix, and perform betweenness centrality analysis on the generated network; The expression of betweenness centrality is as follows: In the above formula, C b (i) represents the betweenness centrality of the node, σ st is the number of shortest paths from node s to node t; σ st (i) is the number of shortest paths from node s to node t that pass through node i.

10. According to the method for identifying safety risks of offshore bridge construction based on text mining and NK-SNA model according to claim 1, step 7) specifically comprises the following steps: The risk factor coupling value obtained in step 5) is used as the correction coefficient to correct the betweenness centrality value obtained in step 6) to obtain the corrected betweenness centrality value. The correction formula is as follows: C b '(i)=T i ×C b (i) (7) In the above formula, C b '(i) represents the corrected betweenness centrality; T i Indicates the risk coupling value of the coupling mode corresponding to the risk factor; C b (i) represents the betweenness centrality of the node; By comparing the betweenness centrality before and after correction, the risk factors whose values ​​are amplified are key risk factors, and the risk factors whose values ​​are reduced are secondary risk factors.

Images