Highway construction safety risk management system

Abstract

The invention relates to the technical field of risk analysis, in particular to an expressway construction safety risk management system, which comprises a data acquisition and time synchronization module for acquiring worker position information, sorting according to timestamps, generating real-time data streams, classifying and standardizing the real-time data streams, and generating classified data streams. According to the method, the risk events are identified more quickly and accurately through classification and analysis of real-time data streams, prediction and intervention before accidents occur are allowed, and potential property loss and personal injury are reduced. And the risk level is determined by combining the priority of the event and the field condition, so that the reasonable allocation of resources and the scientificity of priority ranking are ensured, and the emergency response is more targeted and effective. The risk level is compared with the safety standard, so that the risk control strictness is enhanced, the operation and cost benefits are optimized, and the adaptability and the foresight are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of risk analysis, and in particular to a highway construction safety risk management system. Background Art

[0002] The field of risk analysis technology focuses on identifying, evaluating, and mitigating potential risks that may arise in different environments and situations. This technical field combines statistics, applied mathematics, and decision science to predict and manage risks through quantitative and qualitative methods. The highway construction safety risk management system is to identify and manage safety risks that may be encountered during highway construction.

[0003] However, in fast-changing environments such as highway construction, the rapid changes in environmental and construction conditions make it difficult for existing technologies to effectively prevent new or sudden risks. In addition, there is a lack of dynamic adjustment capabilities when determining risk levels and priority rankings. Once the actual situation changes, the established plans are difficult to meet actual needs, resulting in waste of resources or slow response. This static risk assessment model and slow update mechanism often make it impossible to make quick and effective decisions in emergency situations, increasing the risk of safety accidents and potential losses. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a highway construction safety risk management system.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A highway construction safety risk management system comprises: The data collection and time synchronization module collects the worker location information, sorts it according to the timestamp, generates a real-time data stream, and classifies and standardizes the real-time data stream to generate a classified data stream; An event processing module receives the classified data stream, performs event pattern recognition, detects potential risk events, and generates risk identification results; analyzes the events in the risk identification results, determines event priority and impact range, and generates event priority results; A safety risk assessment module, using the event priority results and combining the on-site conditions to determine the risk level and generate a preliminary risk level; comparing the preliminary risk level with the safety standards and generating a comprehensive risk assessment result; The early warning and response scheduling module formulates an emergency plan based on the comprehensive risk assessment results and generates plan formulation results; and adjusts the construction site and personnel layout according to the plan formulation results.

[0006] Preferably, the steps of acquiring the real-time data stream are: Install positioning sensors to collect real-time location information of workers in the highway construction area to obtain unsorted worker location data; Based on the unsorted worker location data, the data is sorted using timestamps to obtain a data stream arranged in chronological order, thereby generating a sorted data stream; Based on the sorted data stream, time gaps and data duplications are removed to obtain a real-time data stream.

[0007] Preferably, the steps of acquiring the classified data stream are: Receiving the real-time data stream, removing noise and outliers, verifying data integrity using a timestamp, and obtaining a preprocessed data stream; Based on the preprocessed data stream, a weighted normalized score is calculated for each data point using the following formula: ; in, Representative The original value of the data point, represents the average value of the preprocessed data stream, represents the standard deviation of the preprocessed data stream, is the adjustment coefficient, is the weighted standardized score; Based on the weighted normalized scores, a clustering algorithm is used to classify the preprocessed data stream to generate a classified data stream.

[0008] Preferably, the steps for obtaining the risk identification result are: Receiving the classified data stream, screening data points of potential risk indicators, and obtaining a potential risk data set; Based on the potential risk data set, the risk score of each data point is calculated using the following formula: ; in, Representative The characteristic value of the data point, represents the average feature value of the potential risk data set, Represents the characteristic standard deviation of the potential risk data set, is the nonlinear adjustment coefficient, Score the risk; Based on the risk scores, risk identification results are generated by comparing each score with the preset risk threshold.

[0009] Preferably, the steps for obtaining the event priority result are: Based on the risk identification results, the comprehensive priority score of each event is calculated using the following formula: ; in, For events The priority score of For events The risk score of For events The scope of influence, and are the linear weighting coefficients of risk score and impact scope, respectively. is the nonlinear adjustment coefficient, Thresholds for scoring risks; Based on the priority scores, all events are sorted, the processing order of each event is determined, and an event priority result is generated.

[0010] Preferably, the steps for obtaining the preliminary risk level are: receiving the event priority results, analyzing the on-site conditions to determine the environmental factors of each event, and obtaining an environmental adjustment data set; Based on the environmental adjustment data set, the adjusted risk level of each event is calculated using the following formula: ; in, For events The adjusted risk level, For events The priority score of For events The impact value of environmental factors, and are the sensitivity coefficient and threshold of environmental factors respectively; Based on the adjusted risk level, all events were reanalyzed to generate preliminary risk levels.

[0011] Preferably, the steps for obtaining the comprehensive risk assessment results are: Summarize the preliminary risk level data, collect current construction safety standards, and form a safety standard reference data set; Based on the safety standard reference data set, the preliminary risk level of each event is compared, and the consistency or deviation between each risk level and the corresponding safety standard threshold is analyzed to obtain a risk level comparison analysis result; Based on the comparative analysis results of the risk levels, intervention points and priorities are determined according to site conditions and safety standards to obtain a comprehensive risk assessment result.

[0012] Preferably, the steps for obtaining the plan formulation result are: Receive the comprehensive risk assessment results, formulate an emergency plan, plan resource allocation, personnel evacuation routes and emergency contact procedures, and obtain a preliminary emergency plan; The preliminary emergency plan is reviewed and simulated to test the feasibility of the plan, determine the response strategy and action plan for each risk point, and obtain the results of the plan formulation.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, the classification and analysis of real-time data streams make the identification of risk events faster and more accurate, allowing prediction and intervention before accidents occur, reducing potential property losses and personal injuries. The risk level determined in combination with the priority of the event and the on-site situation ensures the rational allocation of resources and the scientific nature of priority sorting, making emergency response more targeted and effective. The comparison of risk levels with safety standards enhances the strictness of risk control, optimizes operation and cost-effectiveness, and improves adaptability and foresight. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0016] See also Figure 1 The present invention provides a technical solution: a highway construction safety risk management system comprising: The data collection and time synchronization module collects the worker location information, sorts it according to the timestamp, generates a real-time data stream, and classifies and standardizes the real-time data stream to generate a classified data stream; An event processing module receives the classified data stream, performs event pattern recognition, detects potential risk events, and generates risk identification results; analyzes the events in the risk identification results, determines event priority and impact range, and generates event priority results; A safety risk assessment module, using the event priority results and combining the on-site conditions to determine the risk level and generate a preliminary risk level; comparing the preliminary risk level with the safety standards and generating a comprehensive risk assessment result; The early warning and response scheduling module formulates an emergency plan based on the comprehensive risk assessment results and generates plan formulation results; and adjusts the construction site and personnel layout according to the plan formulation results.

[0017] The steps of acquiring the real-time data stream are: Install positioning sensors to collect real-time location information of workers in the highway construction area to obtain unsorted worker location data; Based on the unsorted worker location data, the data is sorted using timestamps to obtain a data stream arranged in chronological order, thereby generating a sorted data stream; Based on the sorted data stream, time gaps and data duplications are removed to obtain a real-time data stream.

[0018] Specifically, positioning sensors are installed to collect real-time location information of workers in the highway construction area. First, the installation positions of all sensors are determined to maximize the coverage and ensure that there are no blind spots. The location information is accompanied by a timestamp. These timestamps are calibrated through the built-in synchronization clock to eliminate time errors between devices and obtain unsorted worker location data.

[0019] Based on the unsorted worker location data, the data is sorted based on the timestamp. First, the correctness of the timestamp of each data is checked, and the data with wrong timestamp is eliminated. Then, all the data are sorted according to the timestamp.

[0020] Based on the sorted data stream, time gaps and data duplicates are removed to obtain a real-time data stream. First, the time stamps in the data stream are analyzed, the time gap positions are identified and recorded, and these gaps are filled using linear interpolation to ensure data continuity. For duplicates in the data, only the data items that appear for the first time are retained. Such processing not only reduces the amount of data storage, but also improves data processing speed and efficiency to obtain a real-time data stream.

[0021] The steps of obtaining the classified data stream are: Receiving the real-time data stream, removing noise and outliers, verifying data integrity using a timestamp, and obtaining a preprocessed data stream; Based on the preprocessed data stream, a weighted normalized score is calculated for each data point using the following formula: ; in, Representative The original value of the data point, represents the average value of the preprocessed data stream, represents the standard deviation of the preprocessed data stream, is the adjustment coefficient, is the weighted standardized score; Based on the weighted normalized scores, a clustering algorithm is used to classify the preprocessed data stream to generate a classified data stream.

[0022] Specifically, by scanning the real-time data stream point by point and marking the collection time of the data items, data segments with stable collection frequency and low data value fluctuation range within a fixed period are selected, the values ​​in each data segment are read into an independent record table, and the values ​​in the record table are compared individually. For data items that exceed the maximum deviation threshold obtained by empirical statistics in advance (the threshold is selected by the average value obtained from the distribution statistics of at least 1,000 historical data, and the upward deviation does not exceed 15% of the average value), an elimination operation is performed, the data after the elimination operation is rearranged, and the recording time intervals between the data after the arrangement are checked in a In the time series comparison table that has been manually verified, if there is a time misorder or missing, the positioning timestamps recorded in the independent text file in the original acquisition order are compared one by one, and the abnormal data items are deleted from the table and filled with linear interpolation using the interval values ​​of adjacent data segments. For data items with slight deviations but not exceeding the threshold, the offset correction is based on the data item value minus the median obtained by subtracting the median of 200 adjacent data. The corrected data list is reorganized and the final value is recorded in the independent record table in the same way until all data items are processed, and finally a set of preprocessed data sequences that have been cleaned and checked are obtained.

[0023] The benefit of the formula is that it introduces an absolute deviation term during the normalization process , through the additional amplification of numerical deviations, data points far from the mean are made more prominent in subsequent classifications, The parameter acquisition steps are to use positioning sensors to collect 2000 location data points in the highway construction area for 7 consecutive days, and record the location value of each data point in meters. The parameter acquisition step is to calculate the arithmetic mean of all 2000 position values. , The step of obtaining the parameters is to calculate the standard deviation of the same batch of data. , Parameters are different by comparing Selection of changing features of weighted standardized score distribution when taking values , this value can moderately amplify the deviation value in the range of 0 to 1000 meters.

[0024] Calculation process: Take a single data point location value For example, substitute the parameters , , ,but:

[0025]

[0026]

[0027]

[0028] This result indicates that the current data point has a high deviation value relative to the mean, which means that this data point will be regarded as a unique data sample in the subsequent classification step. This score can be used to identify points that deviate significantly from the mean in the overall distribution, providing a reference for subsequent classification grouping.

[0029] Based on a set of calculated weighted standardized scores based on the starting position, sort the items in the numerical sequence from small to large, mark the position of each score along the horizontal axis, and divide the position sequence into several equal-length intervals according to the established data index table. The score in each interval is compared by numerical comparison to determine the distribution density. If the data distribution in a certain interval is too concentrated, the score sequence in the interval is divided again, and the number and distribution of the divided sub-intervals are recorded in an independent table. The number of target categories n initially defined is selected from the table. The n value is obtained through preliminary actual measurement and analysis of at least 500 similar data distributions to ensure that n is within the fixed value range of 5 to 10. According to the n value, n dividing points are calibrated in the sequence, and the numerical difference between each data point and the adjacent dividing point is checked item by item. For data points that exceed the fixed difference threshold determined by the statistics of 200 historical data (this threshold is 2 in the range of 0 to 5), they are classified into the classification group represented by the corresponding dividing point, and the classification is repeated until all data points are classified into a classification group, eventually forming a set of clearly divided classification result data list.

[0030] The steps for obtaining the risk identification results are: Receiving the classified data stream, screening data points of potential risk indicators, and obtaining a potential risk data set; Based on the potential risk data set, the risk score of each data point is calculated using the following formula: ; in, Representative The characteristic value of the data point, represents the average feature value of the potential risk data set, Represents the characteristic standard deviation of the potential risk data set, is the nonlinear adjustment coefficient, Score the risk; Based on the risk scores, risk identification results are generated by comparing each score with the preset risk threshold.

[0031] Specifically, access the classified data stream, mark the eigenvalue corresponding to each data point in a single column, put at least 1000 data point eigenvalues ​​collected in consecutive days into a matrix, extract values ​​by scanning row by row, and compare each data value with a manually annotated reference feature file. The file records the characteristic distribution intervals corresponding to the unconventional areas confirmed by on-site manual investigation in the past period of time, and each interval is represented by a numerical range. The reference feature file is obtained by comparing more than 500 data in the actual construction site, and the threshold is obtained by the median of these comparison results floating upward by no more than 15% of the median. The numerical intervals given in the reference file are compared from the matrix data one by one, and the data values ​​matching the reference intervals are marked in an independent list. The timestamp of each matching item is recorded in the synchronous record text, and then the timestamps are sorted in the synchronous record text. The marked data values ​​of the continuous time period are selected from the sorting results to form a new numerical column, and each data point in these numerical columns is marked in a separate table in a single column. The table is specially used for the arrangement of potential risk indicator data points. When the table content is recorded, a potential risk data set is generated.

[0032] The benefit of the formula is that it converts the deviation of the original data into a more distinguishable risk score by logarithmizing the difference between the eigenvalue and the mean value, introduces the standard deviation as a normalization scale, and uses a nonlinear adjustment coefficient to improve sensitivity to extreme values.

[0033] The parameter acquisition step is to continuously monitor the characteristic values ​​of potential risk points in the construction area 24 hours a day in the past 7 days, detect the vertical offset of workers relative to the safety baseline, collect a total of 1,200 characteristic value records, and average these characteristic values ​​to obtain Parameters, statistically ; The parameters are calculated by the standard deviation of the same batch of data. ; The parameters were selected by using a nonlinear curve fitting method in the analysis of historical data over the past 3 months, and comparing multiple values ​​in the range of 0.5 to 5.0. , the risk identification of extreme values ​​in this range is more obvious.

[0034] Calculation process: Take a single data point For example, bring in , , :

[0035]

[0036]

[0037]

[0038]

[0039] final ; This result shows that the current data point is not extremely deviated from the overall average characteristic value. Values ​​below 1.0 indicate a lower risk. A value greater than 2.0 indicates that the data point is seriously deviated, providing a basis for marking it as a high-risk point in subsequent risk identification.

[0040] At the beginning, the previous risk scoring result list is independently marked, and these scoring results are read into independent text records one by one, and the scoring value of each data point in the text record is arranged in ascending order. For each scoring value in the arrangement sequence, a threshold parameter table generated by historical event statistics is matched respectively. The threshold parameter table is obtained by the statistics of the actual event feature values ​​recorded in the past 90 days. The specific threshold is calculated by the median of the extreme value interval of multi-day records, and then a step value obtained by manual experience is added to this median value range. This step value is determined by analyzing the score difference of 200 real events. In the matching process, each scoring value is checked for its numerical difference with the corresponding record in the threshold parameter table, and the scoring items that meet the condition greater than the threshold expression are picked out from the record and marked. After marking, these marked items are summarized in an independent output list without further processing of the scoring distribution characteristics and parameter adjustment steps. The scoring marking results of each item in the list are presented in sequence in a final result document as the output of the risk identification result.

[0041] The steps for obtaining the event priority result are: Based on the risk identification results, the comprehensive priority score of each event is calculated using the following formula: ; in, For events The priority score, For events The risk score of For events The scope of influence, and are the linear weighting coefficients of risk score and impact scope, respectively. is the nonlinear adjustment coefficient, Thresholds for scoring risks; Based on the priority scores, all events are sorted, the processing order of each event is determined, and an event priority result is generated.

[0042] Specifically, the formula is beneficial in that it linearly weights the risk score and the impact range, and introduces a and The sigmoid function controlled by the parameters produces a nonlinear enhancement on the priority score when the event risk score exceeds the threshold, thereby highlighting the key events.

[0043] Parameters, after data cleaning and statistical analysis, the risk score of event i is extracted from the 400 identified event data ; The parameter acquisition step is to use a laser ranging and range demarcation instrument to measure the affected area in the area where event i is located, and obtain it by actual measurement of a local area with a radius of no more than 20 meters. , this value comes from multiple field tests and records; and The parameter acquisition step is to perform linear regression and residual analysis on the data of the past three months, associate the event risk score with the actual subsequent loss degree and select ; Compare the scope of event impact with the actual corresponding resource input and weight fitting, and finally select ; The parameter acquisition step is to apply a nonlinear fitting method to the score distribution curve and compare multiple The impact of the value on the score growth rate at different thresholds is determined ; The parameter acquisition step is to statistically analyze the risk score distribution in the historical data, select the critical value of the risk score distribution interval, and determine the .

[0044] Calculation process: Take event i as an example: :

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052] The results show that for event i, the priority score is approximately 3.735. A value greater than 1.0 indicates that the risk and impact range of the event have obtained a higher priority through linear and nonlinear combinations. A larger score indicates that the event is more worthy of priority in subsequent processing steps.

[0053] Refer to the numerical column in the event record file for which the previous priority score has been calculated, extract the priority score corresponding to each event in the record file into an independent data row, and mark the event number in the first column of the data row. After the event number is associated with the corresponding priority score one-to-one, all priority scores in the data row are arranged in order from large to small. In order to clearly define the threshold and grouping reference line in the sorting process, the median priority value extracted from the monitoring data of the past 30 days is used as the statistical basis, and the median priority value is recorded as M. If the score of an event exceeds M, it is recorded in one type of data list. If the event score is lower than M, it is recorded in another type of data list. The two lists are recorded independently in two independent files. , and scan the data rows in the two files again. For each event record in the scanning process, read the additional information parameters related to the event. The parameters come from a specific event characteristic record text accumulated for 5 days. The additional information parameters are recorded at the end of the line that matches the event by corresponding the event number with the additional information parameter number. Then, the additional information parameters at the end of the line are compared numerically, and events that are higher than the fixed threshold value obtained previously through statistics of no less than 100 event data (the threshold value is obtained by fine-tuning the average value of the distribution range of the event additional information parameters, for example, increasing the average value by 10%) are marked in an independent list. The list is finally output in text form and the entire sorting and numbering process ends.

[0054] The steps for obtaining the preliminary risk level are: receiving the event priority results, analyzing the on-site conditions to determine the environmental factors of each event, and obtaining an environmental adjustment data set; Based on the environmental adjustment data set, the adjusted risk level of each event is calculated using the following formula: ; in, For events The adjusted risk level, For events The priority score, For events The impact value of environmental factors, and are the sensitivity coefficient and threshold of environmental factors respectively; Based on the adjusted risk level, all events were reanalyzed to generate preliminary risk levels.

[0055] Specifically, access the first column information of each event record line read from the event priority result file, decompose the information into event numbers and corresponding on-site basic parameters, correspond the event numbers to the environmental characteristic data recorded in the on-site collection documents one by one, form a two-dimensional data table for each corresponding result, and extract the temperature, humidity, air pressure and day and night light intensity values ​​involved in the corresponding event from the file one by one for the environmental parameters of each record in the table, based on the on-site specific influencing factor file collected in the early stage. These values ​​are obtained through time series measurement of at least 100 data points, and are recorded in an independent text list for verification. All extracted parameters are stored in the data table by line, and the distribution of these parameters is scanned item by item, and selected from them. Record items that exceed the upper temperature limit or lower humidity limit determined by empirical statistics. The empirical statistics are based on the calculation of the average value of the temperature distribution range in the past 30 days of monitoring and an additional floating range of no more than 15% is added to the average value. If the temperature value is higher than this floating range, it is marked as a high temperature event. Similarly, the humidity is calculated based on the same period data and a fixed threshold is set. Once the humidity value is lower than the threshold, it is marked as a low humidity event. For the parameters of the marked event, the pressure and illuminance distribution are compared in the reference file, and the illuminance data is recorded in the range of 0 to 1000 lux according to the intensity value provided by the actual measuring instrument. Compared with the middle value calculated in advance, events above the middle value are recorded as records with strong light, and then these marking information are stored in a unified record list for summary.

[0056] The formula is useful in that the priority score Based on the environmental factors affecting the value With threshold The difference between them is nonlinearly mapped. Significant deviation When the risk level is adjusted, it will have a significant amplifying or reducing effect, thus reflecting the sensitive regulation of environmental factors on the risk level of the event.

[0057] Parameters, priority calculation is performed on 200 events for 7 consecutive days, and the priority score of event i is recorded as ; The parameter acquisition step is to obtain the environmental parameters of the day in the area where event i is located through thermometers, hygrometers and illuminance meters, and use standardized methods to convert multiple parameters into impact value ranges from 0 to 10. ; The parameter acquisition step is to perform nonlinear fitting on the environmental factors affecting the data of hundreds of historical events and compare different The influence of the value on the steepness of the score curve is shown in is the selected value; The step of obtaining the parameters is to count the impact values ​​of environmental factors in the historical distribution and select the middle value of the densely distributed interval as ; Calculation process: Taking event i as an example, , , , :

[0058]

[0059]

[0060]

[0061]

[0062] The results show that after considering the impact of environmental factors, the risk level of event i is about 2.09. If it is greater than 2.0, it means that the risk of the event is high under the current environmental conditions. Values ​​below 1.0 indicate low risk, while A value well above 3.0 means that the event is very significant under the current environmental conditions.

[0063] Refer to the preliminary risk level data table generated in the early stage, indicate the event number in the first column of the data table, and search the corresponding historical event environment records in the attached files. For each event number, read multiple environmental numerical parameters in the past 24 hours from the attached files one by one, divide these parameters into intervals with a fixed step size and record them in an independent table. For the uneven distribution of parameters, perform statistical calculations on at least 100 event records and set a reference numerical dividing point. The dividing point is determined by floating 10% upward from the median value of the historical distribution. If a parameter exceeds the dividing point, the event is marked in an independent list, and the list is sorted in chronological order. The first few event records that occurred most recently in the sorting results are extracted, and the parameters in these records are compared twice. The results after the secondary comparison are integrated into the final preliminary risk level output list.

[0064] The steps for obtaining the comprehensive risk assessment results are: Summarize the preliminary risk level data, collect current construction safety standards, and form a safety standard reference data set; Based on the safety standard reference data set, the preliminary risk level of each event is compared, and the consistency or deviation between each risk level and the corresponding safety standard threshold is analyzed to obtain a risk level comparison analysis result; Based on the comparative analysis results of the risk levels, intervention points and priorities are determined according to site conditions and safety standards to obtain a comprehensive risk assessment result.

[0065] Specifically, the collected event risk levels are compared in detail with the safety thresholds. During the comparison process, the risk level of each event is analyzed to see if it exceeds the set safety thresholds. These thresholds are adjusted based on industry standards and past accident analysis to ensure that all potential danger levels are covered. After the analysis is completed, the risk level comparison analysis results are generated.

[0066] Based on the comparison of risk levels with safety standards, risk ranking and classification are performed. All events are ranked from high to low risk, and the urgency and priority of each event are classified. The ranking process is based on the risk level of the event and the detailed intervention thresholds in the safety standards. These intervention thresholds are based on the latest safety research and past lessons learned, including detailed response strategies for different risk levels.

[0067] The steps for obtaining the plan formulation results are as follows: Receive the comprehensive risk assessment results, formulate an emergency plan, plan resource allocation, personnel evacuation routes and emergency contact procedures, and obtain a preliminary emergency plan; The preliminary emergency plan is reviewed and simulated to test the feasibility of the plan, determine the response strategy and action plan for each risk point, and obtain the results of the plan formulation.

[0068] Specifically, based on the results of the comprehensive risk assessment, we began to develop emergency plans and allocate necessary rescue resources to each risk area, such as locating the nearest fire station and medical center, ensuring that the rescue path is unobstructed, and taking the shortest path into consideration when developing evacuation routes. At the same time, we established an emergency contact process to ensure that all key personnel can be notified quickly when an incident occurs.

[0069] Conduct field audits and simulation drills on the preliminary emergency plan to ensure its practical feasibility. During the drills, pay special attention to the actual traffic conditions of the evacuation routes and the response speed of the emergency contact process. Adjustments are made to any problems found, such as potential blockages in the evacuation routes or delays in the contact process, to enhance the practicality and efficiency of the plan. After multiple drills, the plan is modified based on feedback to determine the response strategy and action plan for each risk point.

Claims

1. A highway construction safety risk management system, characterized in that: The system comprises: The data collection and time synchronization module collects the worker location information, sorts it according to the timestamp, generates a real-time data stream, and classifies and standardizes the real-time data stream to generate a classified data stream; An event processing module receives the classified data stream, performs event pattern recognition, detects potential risk events, and generates risk identification results; analyzes the events in the risk identification results, determines event priority and impact range, and generates event priority results; A safety risk assessment module, using the event priority results and combining the on-site conditions to determine the risk level and generate a preliminary risk level; comparing the preliminary risk level with the safety standards and generating a comprehensive risk assessment result; The early warning and response scheduling module formulates an emergency plan based on the comprehensive risk assessment results and generates a plan formulation result; and adjusts the construction site and personnel layout according to the plan formulation result; The steps for obtaining the event priority result are: Based on the risk identification results, the comprehensive priority score of each event is calculated using the following formula: ； in, For events The priority score, For events The risk score of For events The scope of influence, and are the linear weighting coefficients of risk score and impact scope, respectively. is the nonlinear adjustment coefficient, Thresholds for scoring risks; Based on the priority scores, all events are sorted, the processing order of each event is determined, and an event priority result is generated.

2. The highway construction safety risk management system according to claim 1 is characterized in that: The steps of acquiring the real-time data stream are: Install positioning sensors to collect real-time location information of workers in the highway construction area to obtain unsorted worker location data; Based on the unsorted worker location data, the data is sorted using timestamps to obtain a data stream arranged in chronological order, thereby generating a sorted data stream; Based on the sorted data stream, time gaps and data duplications are removed to obtain a real-time data stream.

3. The highway construction safety risk management system according to claim 1 is characterized in that: The steps of obtaining the classified data stream are: Receiving the real-time data stream, removing noise and outliers, verifying data integrity using a timestamp, and obtaining a preprocessed data stream; Based on the preprocessed data stream, a weighted normalized score is calculated for each data point using the following formula: ； in, Representative The original value of the data point, represents the average value of the preprocessed data stream, represents the standard deviation of the preprocessed data stream, is the adjustment coefficient, is the weighted standardized score; Based on the weighted normalized scores, a clustering algorithm is used to classify the preprocessed data stream to generate a classified data stream.

4. The highway construction safety risk management system according to claim 1 is characterized in that: The steps for obtaining the risk identification results are: Receiving the classified data stream, screening data points of potential risk indicators, and obtaining a potential risk data set; Based on the potential risk data set, the risk score of each data point is calculated using the following formula: ； in, Representative The characteristic value of the data point, represents the average feature value of the potential risk data set, Represents the characteristic standard deviation of the potential risk data set, is the nonlinear adjustment coefficient, Score the risk; Based on the risk scores, risk identification results are generated by comparing each score with the preset risk threshold.

5. The highway construction safety risk management system according to claim 1 is characterized in that: The steps for obtaining the preliminary risk level are: receiving the event priority results, analyzing the on-site conditions to determine the environmental factors of each event, and obtaining an environmental adjustment data set; Based on the environmental adjustment data set, the adjusted risk level of each event is calculated using the following formula: ； in, For events The adjusted risk level, For events The priority score, For events The impact value of environmental factors, and are the sensitivity coefficient and threshold of environmental factors respectively; Based on the adjusted risk level, all events were reanalyzed to generate preliminary risk levels.

6. The highway construction safety risk management system according to claim 1 is characterized in that: The steps for obtaining the comprehensive risk assessment results are: Summarize the preliminary risk level data, collect current construction safety standards, and form a safety standard reference data set; Based on the safety standard reference data set, the preliminary risk level of each event is compared, and the consistency or deviation between each risk level and the corresponding safety standard threshold is analyzed to obtain a risk level comparison analysis result; Based on the comparative analysis results of the risk levels, intervention points and priorities are determined according to site conditions and safety standards to obtain a comprehensive risk assessment result.

7. The highway construction safety risk management system according to claim 1 is characterized in that: The steps for obtaining the plan formulation results are as follows: Receive the comprehensive risk assessment results, formulate an emergency plan, plan resource allocation, personnel evacuation routes and emergency contact procedures, and obtain a preliminary emergency plan; The preliminary emergency plan is reviewed and simulated to test the feasibility of the plan, determine the response strategy and action plan for each risk point, and obtain the results of the plan formulation.

Images