CN118428735A

CN118428735A - Engineering project and large-risk engineering integrated safety management system

Info

Publication number: CN118428735A
Application number: CN202410602887.5A
Authority: CN
Inventors: 杨世伟; 成浪; 郭林飞; 谭秋林; 吴强; 刘湘远; 谭鑫龙; 席星; 魏江坤; 王强; 袁内蒙; 何煜民
Original assignee: China Construction First Group Corp Ltd
Current assignee: China Construction First Group Corp Ltd
Priority date: 2024-05-15
Filing date: 2024-05-15
Publication date: 2024-08-02

Abstract

The invention relates to the technical field of safety engineering, in particular to an engineering project danger engineering integrated safety management system. The modules of the system include the following functions: constructing an engineering risk prediction model; carrying out risk prediction on a dangerous large project according to the project risk prediction model to obtain project potential risk data; performing effective risk prevention list processing according to the engineering potential risk data to obtain effective prevention list data; performing on-site hidden danger real-time monitoring processing according to the effective prevention list data, and performing risk trend prediction processing to obtain risk trend prediction data; carrying out weekly control early warning processing according to the risk trend prediction data to obtain weekly risk control early warning data; and performing performance punishment on responsible persons through weekly risk management and control early warning data, and performing project risk marketing. According to the invention, risk management and control are carried out through three dimensions in advance, in the event and after the event, so that effective prevention and management of the safety risk of the dangerous large engineering project are realized.

Description

Engineering project and large-risk engineering integrated safety management system

Technical Field

The invention relates to the technical field of safety engineering, in particular to an engineering project danger engineering integrated safety management system.

Background

With the development of social economy and the progress of science and technology, the scale of engineering projects is continuously enlarged, the engineering construction environment is increasingly complex, engineering construction safety faces more and more challenges, and particularly in the field of dangerous engineering, such as nuclear power plants, chemical plants and the like, the occurrence of safety accidents generally brings about huge environmental and personnel injuries. In recent years, large and above security accidents of house construction and municipal infrastructure engineering still occur, and the accidents cause serious life and property loss and adverse social influence, and the reasons of the accidents are related to special construction schemes of dangerous engineering. Therefore, the reinforced-risk large engineering management has extremely important significance for effectively restraining the occurrence of safety accidents and guaranteeing the safety and stability of the construction process. However, the conventional engineering project dangerous large engineering safety management is often in a scattered and single management mode, construction risk management and control measures are improper, emergency measures are not strong in pertinence, even problems such as standard regulation violation occur, and therefore the safety condition of the project cannot be comprehensively mastered, and the safety management work of a construction site cannot be effectively guided.

Disclosure of Invention

Based on the above, the invention provides an engineering project risk engineering integrated safety management system to solve at least one of the above technical problems.

In order to achieve the above purpose, an engineering project danger large engineering integrated safety management system comprises:

The risk identification evaluation module is used for acquiring historical engineering case data; constructing an engineering risk prediction model based on the historical engineering case data; carrying out risk prediction on a dangerous large project according to the project risk prediction model to obtain project potential risk data; carrying out engineering risk scale assessment according to engineering potential risk data to obtain risk scale data of the dangerous large engineering;

The multi-dimensional risk index construction module is used for constructing a multi-dimensional risk index system according to engineering potential risk data; performing risk matrix evaluation processing on the multidimensional risk index system by using risk scale data of the risk large engineering to obtain risk evaluation grade data; performing effective prevention list approval according to the risk assessment grade data to obtain effective prevention list data;

the daily management and control module is used for distributing the daily risk management and control task to the effective prevention list data to obtain daily risk management and control task data; performing on-site hidden danger real-time monitoring processing according to daily risk management and control task data to obtain real-time hidden danger monitoring data; performing management and control completion rate analysis on the real-time hidden danger monitoring data through daily risk management and control task data to obtain risk management and control completion rate data;

the risk supervision processing module is used for carrying out hidden trouble context fusion processing on the real-time hidden trouble monitoring data to obtain hidden trouble monitoring fusion data; carrying out dynamic risk clustering processing according to the hidden danger monitoring fusion data to obtain dynamic risk clustering data; carrying out risk trend prediction processing based on the dynamic risk cluster data to obtain risk trend prediction data;

The weekly monitoring module is used for carrying out management and control quality calculation on the risk trend prediction data to obtain hidden danger point management and control quality data; carrying out weekly control early warning processing on responsible persons through hidden danger point control quality data to obtain weekly risk control early warning data;

The project early warning treatment module is used for carrying out performance punishment on responsible persons through weekly risk management and control early warning data to obtain a risk management and control punishment strategy; and carrying out risk marketing item processing based on a risk management and control punishment strategy to obtain item risk marketing item data.

By acquiring the historical engineering case data, various potential risk factors and occurrence frequencies thereof can be identified, which is helpful for project teams to know the risks existing in the engineering project planning stage and prepare precaution and control measures in advance. By analyzing the historical engineering case data, certain rules and trends can be found, and a reliable risk prediction model can be built. Such a model can predict the type and probability of risk that occurs by comparing and analyzing the conditions of the current engineering project. Based on the established risk prediction model, the specific situation of the current engineering project can be predicted, and corresponding risk data can be obtained. By evaluating and analyzing the engineering potential risk data, the risk scale of the engineering project, namely the degree of loss and the influence range can be determined. By constructing a multidimensional risk index system according to engineering potential risk data, risk factors can be systematically classified and generalized, so that comprehensive multidimensional risk assessment indexes are formed. By combining the risk scale data of the dangerous large project with the multidimensional risk index system, the risk of the project can be comprehensively evaluated. Through the risk matrix evaluation processing, the influence degree and possibility of different risk factors can be quantified, and corresponding risk evaluation grade data can be obtained. Based on the risk assessment grade data, risk items with higher influence degree and higher possibility can be determined, and corresponding counter measures and prevention lists are further determined. By performing daily risk management and control task dispatch on the effective prevention list data, project teams are ensured to know the risk items needing to be focused on the day in time, and the task dispatch can enable project managers and field staff to clearly know the daily risk tasks needing to be processed. Based on daily risk management and control task data, project teams can monitor and process real-time hidden dangers on site, and various potential hidden dangers and risks existing in engineering projects can be timely found through such real-time monitoring. The real-time hidden danger monitoring data and daily risk management and control task data are compared and analyzed, so that the completion condition of risk management and control work can be estimated. Hidden trouble context fusion processing is carried out on real-time hidden trouble monitoring data, and hidden trouble data of different time points and different monitoring dimensions can be integrated and fused. And dividing hidden danger data according to the similarity, and identifying hidden danger points of different categories and aggregation trends of the hidden danger points. By analyzing and mining the dynamic risk cluster data, the development trend and the evolution path of hidden danger points can be found. The effectiveness of risk management and control work can be evaluated and monitored by carrying out management and control quality calculation on risk trend prediction data. Based on the hidden danger point management and control quality data, the performance of the responsible person in the risk management and control work can be evaluated and monitored. A set of scientific and reasonable risk management and control punishment strategies can be established by carrying out performance punishment on responsible persons through weekly risk management and control early warning data. The strategy can evaluate the performance and performance of the responsible person in the risk management and control work, and proper punishment measures are adopted for the responsible person who fails to effectively fulfill the responsibilities, so that the responsible person is stimulated to participate in the risk management and control work more seriously, and the work quality and efficiency are improved. And according to the established risk management and control punishment strategy, marketing items are processed on the risks which are solved by taking effective measures. The processing can ensure the timely follow-up and implementation of risk management and control work, reduce accumulation and retention of potential risks and guarantee the smooth proceeding of projects. Therefore, the integrated safety management system for the engineering project and the dangerous large engineering carries out potential risk prediction on the dangerous large engineering project through three dimensions of advance, in-process and after-process, and establishes a risk relief strategy through a multi-dimensional risk index system. Daily risk management and control task processing is carried out according to a risk relief strategy, responsible person warning processing is carried out by considering weekly monitoring data, corresponding performance punishment is carried out according to warning information, the management responsibility of the floor is enhanced through quantitative assessment indexes and index completion and performance assessment hooks, and the safety risk of a danger-controlling large project in the construction process is effectively reduced.

Preferably, the risk identification evaluation module comprises the following functions:

Acquiring historical engineering case data and engineering project document data;

Training an engineering risk prediction model according to the historical engineering case data to obtain an engineering risk prediction model;

carrying out natural language processing on the project document data to obtain target project feature data;

Dividing the dangerous engineering types according to the target engineering characteristic data to obtain dangerous engineering type data;

Carrying out engineering potential risk prediction on the target engineering characteristic data by utilizing an engineering risk prediction model to obtain engineering potential risk data;

And carrying out engineering risk scale assessment on the engineering potential risk data through the dangerous large engineering type data to obtain dangerous large engineering risk scale data.

The invention can establish the basis of comprehensive understanding and deep analysis of engineering projects by legally collecting historical engineering case data and engineering project document data. Historical engineering case data may help identify common risk points and potential problems, while engineering project document data contains project details and features. By utilizing the historical engineering case data, an engineering risk prediction model can be established, so that the risk of a target engineering project is predicted and estimated. Such a model can predict the risk type and probability faced by future engineering projects by analyzing various factors and events in the historical cases. By performing natural language processing on the project document data, various features and key information of the target project can be extracted from the project document data. By analyzing and dividing the target engineering characteristic data, the dangerous engineering type of the engineering project can be determined. Based on the established project risk prediction model, the target project characteristic data is analyzed and predicted, so that the potential risk situation of the project can be obtained, and the project team can be helped to find potential safety hazards and risk points in time through the prediction. Based on the determined risk large project type and project potential risk data, the risk scale of the project can be estimated and quantified, and the estimation can help project teams comprehensively know the risk degree and influence scope of the project.

Preferably, the training of the engineering risk prediction model according to the historical engineering case data is performed, including:

automatically preprocessing the historical engineering case data to obtain preprocessed historical engineering case data;

carrying out characteristic engineering processing according to the preprocessed historical engineering case data to obtain the historical engineering case characteristic data;

Performing recursive feature elimination processing on the historical engineering case feature data to obtain historical critical engineering feature data;

constructing a feature library based on the historical critical engineering feature data to obtain a critical engineering feature library;

mining risk factors according to the risk engineering feature library to obtain risk factor feature data;

And performing migration learning on the risk factor characteristic data by using a preset convolutional neural network model, and performing characteristic weight optimization to obtain an engineering risk prediction model.

The invention carries out automatic pretreatment on the historical engineering case data, wherein the automatic pretreatment can comprise the steps of data cleaning, repeated value removal, missing value filling, data transformation and the like so as to ensure the quality and consistency of the data. And carrying out feature engineering processing according to the preprocessing history engineering case data, and extracting features which are significant to modeling tasks from the original data. The method has the advantages that the method can be used for screening out the characteristics with the most influence on modeling tasks by performing recursive characteristic elimination processing on the historical engineering case characteristic data, so that the characteristics with smaller contribution to the modeling can be eliminated, the modeling efficiency and accuracy are improved, the model structure is simplified, and the risk of overfitting is reduced. The construction of the feature library is carried out based on the historical dangerous engineering feature data, and key features in the historical engineering cases can be integrated and generalized by the construction, so that a complete feature library is formed. By mining key features and risk factors in the critical engineering feature library, factors with great influence on the risk of the critical engineering can be found, and the risk sources and characteristics of the critical engineering can be deeply understood through mining. And performing migration learning by using a convolutional neural network model, and applying key features mined from historical engineering cases to a risk prediction task of the target engineering. Through feature weight optimization, the prediction performance and generalization capability of the model can be further improved, and accurate prediction and effective management of engineering risks are realized.

Preferably, the multi-dimensional risk indicator construction module includes the following functions:

constructing a multidimensional risk index system according to engineering potential risk data and target engineering characteristic data;

Carrying out risk factor weight quantitative processing on engineering potential risk data to obtain risk factor weight data;

Performing risk matrix evaluation processing on the multidimensional risk index system by using risk scale data of the risk large engineering based on the risk factor weight data to obtain risk evaluation grade data;

Making a risk relief strategy according to the risk assessment grade data to obtain risk preventive measure list data;

And carrying out effective prevention list approval according to the risk prevention measure list data to obtain effective prevention list data.

According to the invention, through comprehensively analyzing and integrating the engineering potential risk data and the target engineering characteristic data, a multidimensional risk index system can be established, and the index system can comprehensively consider various potential risk factors and engineering characteristics and comprehensively evaluate the risk conditions faced by engineering projects. The importance and contribution degree of each risk factor can be determined by carrying out quantitative analysis and weighting treatment on engineering potential risk data. By utilizing the risk factor weight data and the risk scale data of the dangerous large engineering, the multidimensional risk index system can be comprehensively evaluated and graded, and the evaluation can quantify the influence degree and possibility of various risk factors and convert the influence degree and possibility into corresponding risk evaluation grade data. Based on the risk assessment grade data, risk items with higher influence degree and higher possibility can be determined, and corresponding risk relief strategies and preventive measures are formulated. By examining and approving the risk preventive measure list data, scientificity and implementation of preventive measures can be ensured, the examination and approval process can ensure effective implementation of preventive measures, various potential risks faced by engineering projects are reduced to the greatest extent, and safe implementation of the engineering projects is ensured.

Preferably, the construction of the multidimensional risk index system according to the engineering potential risk data and the target engineering characteristic data includes:

Carrying out historical risk occurrence frequency statistics according to engineering potential risk data, and carrying out risk recurrence probability calculation to obtain a historical risk occurrence probability index;

Carrying out safe production standardized evaluation score calculation according to the target engineering characteristic data to obtain a construction side field management level index;

Performing risk burst emergency bearing capacity assessment on the target engineering characteristic data based on the engineering potential risk data to obtain project emergency bearing capacity indexes;

carrying out historical emergency event cluster analysis according to engineering potential risk data to obtain clustered historical risk event data;

Performing casualty grade calculation of risk event personnel according to the clustered historical risk event data to obtain casualty grade data;

carrying out risk event economic loss evaluation according to the clustered historical risk event data to obtain economic loss grade data;

performing sensitive target influence level processing around the risk event according to the clustered historical risk event data, and simultaneously performing social attention mining to obtain social influence loss data;

Performing fault influence degree calculation according to the clustering historical risk event data to obtain fault influence degree data, wherein the fault influence degree data comprises infrastructure damage influence degree data and project life guarantee interruption influence degree data;

Carrying out weighted average calculation on the probability indexes of the occurrence of the historical risk, the field management level indexes of the construction side and the emergency bearing capacity indexes of the project to obtain risk probability indexes; and carrying out weighted average calculation on the result severity of the casualties, the economic loss, the social influence loss and the fault influence degree to obtain the risk result severity index.

According to the invention, through statistics of the occurrence frequency of the historical risks, the team can quantify the occurrence conditions of different types of risks in the past, and know the relative importance of the risks; the risk recurrence probability calculation can evaluate the recurrence probability of the historical risk event and remind the project team of the repeated risk event, so that corresponding precautions are taken. The security production standardized review score calculation can evaluate the security management level of the constructor from multiple dimensions, including the soundness of a security management system, the development condition of security training, the integrity of security production records and the like. By analyzing the engineering potential risk data, the emergency bearing capacity of the target engineering in the case of facing an emergency risk event can be estimated, and the risk resistance of the project in the case of facing the emergency risk event can be comprehensively known. By carrying out cluster analysis on engineering potential risk data, historical emergency events can be classified and integrated according to similarity. By analyzing and evaluating the clustering historical risk event data, the influence degree of various risk events on casualties can be determined. By analyzing and evaluating the clustered historical risk event data, the degree of economic loss caused by various risk events can be determined. Such rating data can help a project team understand the potential impact of different risk events on project economies, and evaluation of economic loss ratings can be based on quantitative analysis of the extent of damage to engineering projects, facilities, assets, etc. by risk events. By analyzing the clustered historical risk event data, the influence degree of various risk events on surrounding sensitive targets and social attention can be determined, and potential influences of the risk events on the aspects of environment, public safety feeling, social stability and the like can be considered in the mining of social influence loss data. By analyzing and evaluating the clustered historical risk event data, the degree of fault influence caused by various risk events can be determined, and the degree of influence of infrastructure damage on project progress, cost and safety and the degree of influence of project life guarantee interruption on project operation and personnel life can be considered in calculation of the degree of fault influence. The probability of occurrence of the risk event can be comprehensively evaluated by carrying out weighted average calculation on the historical risk occurrence probability index, the construction side field management level index and the project emergency bearing capacity index. The severity of the consequences of the risk event can be comprehensively evaluated by weighted average calculation of casualty grade data, economic loss grade data, social impact loss data and fault impact degree data. The weighted average calculation of the result severity can comprehensively consider the influence degree of factors such as casualties, economic losses, social influence, fault influence and the like on projects, and is beneficial to determining management key points and coping strategies of risk events.

Preferably, the daily management module comprises the following functions:

The effective prevention list data are imported into a daily management and control module, and daily risk management and control task assignment is carried out on the effective prevention list data to obtain daily risk management and control task data;

Performing risk hidden danger check point processing according to daily risk management and control task data to obtain risk monitoring point data;

Performing on-site hidden danger real-time monitoring processing based on the risk monitoring point data to obtain real-time hidden danger monitoring data;

transmitting the real-time hidden danger monitoring data to a daily management and control module for reporting the management and control task to obtain management and control task reporting data;

and carrying out management and control completion rate calculation on management and control task report data through daily risk management and control task data based on the daily management and control module to obtain risk management and control completion rate data.

According to the invention, the effective prevention list data is imported into the daily management and control module, so that a project team can timely know specific tasks needing risk management and control. The daily risk management and control task is dispatched, so that the ordered performance of the risk management and control task can be ensured, and the safe and stable operation of the project is ensured. By processing daily risk management and control task data, a project team can determine risk hidden danger points to be checked and record the risk hidden danger points as risk monitoring point data. Through carrying out real-time monitoring processing to risk monitoring point data, project team can in time discover and record the risk hidden danger condition that exists on the scene, and on-the-spot hidden danger real-time monitoring can in time track and record on-spot risk condition through real-time data acquisition and monitoring system. Through transmitting real-time hidden danger monitoring data to the daily management and control module, project team can report the on-the-spot monitoring condition in time. The project team can accurately evaluate the completion condition of the daily risk management and control task by processing and analyzing the management and control task report data, and the completion condition of the management and control task can be intuitively presented by the risk management and control completion rate data.

Preferably, the risk supervision processing module comprises the following functions:

Performing image frame enhancement processing on the real-time hidden danger monitoring data to obtain hidden danger monitoring image frame data;

Performing image segmentation processing according to the hidden danger monitoring image frame data to obtain hidden danger monitoring image segmentation data;

carrying out hidden danger source object identification on hidden danger monitoring image segmentation data to respectively obtain hidden danger source object characteristic data and monitoring external scene data;

Hidden trouble context fusion processing is carried out on hidden trouble source object feature data through monitoring external scene data, so that hidden trouble monitoring fusion data are obtained;

performing risk management and control level assessment according to hidden danger monitoring fusion data to obtain real-time risk management and control level data; carrying out dynamic risk clustering processing according to the hidden danger monitoring fusion data to obtain dynamic risk clustering data;

and carrying out risk trend prediction processing on the real-time risk management and control level data through the dynamic risk cluster data to obtain risk trend prediction data.

The invention can improve the quality and definition of the hidden danger monitoring image through image frame enhancement processing, so that hidden danger monitoring data are more accurate and reliable. Through image segmentation processing, a target object in the hidden danger monitoring image can be separated from the background, so that the position and the shape of the hidden danger are more clearly visible, and the processing is helpful for accurately positioning the position and the boundary of the hidden danger. Through hidden danger source object identification, hidden danger source objects in the image can be distinguished from other scenes, and characteristic data of the hidden danger source objects can be extracted. Meanwhile, the monitoring external scene data including surrounding environment information related to hidden danger can be acquired. The background and the environment where hidden danger occurs can be more comprehensively known by fusing the monitored external scene data and the hidden danger source object characteristic data. Such fusion processing can help identify the cause and influencing factors of the hidden trouble. Through evaluating hidden danger monitoring fusion data, the level and effect of current risk management and control can be known in time. By carrying out dynamic risk clustering processing on hidden danger monitoring fusion data, risk events with similar characteristics can be subjected to cluster analysis, so that potential risk modes and trends are identified.

Preferably, executing the risk trend prediction processing on the real-time risk management level data by the dynamic risk cluster data includes:

performing association rule mining on the dynamic risk clustering data to obtain dynamic risk clustering data;

Carrying out dynamic risk factor identification according to the dynamic risk cluster data to obtain dynamic risk factor data;

performing Monte Carlo simulation according to the dynamic risk factor data, so as to obtain dynamic risk propagation simulation data;

extracting risk propagation time sequence according to the dynamic risk propagation simulation data to obtain risk propagation time sequence data;

trend feature extraction is carried out based on the risk propagation time sequence data, so that risk trend feature data are obtained;

and carrying out risk trend prediction processing on the real-time risk management and control level data by using the risk trend characteristic data to obtain risk trend prediction data.

By adopting the association rule mining, the invention can find out the relativity and regularity of different dynamic risk events and reveal the association relation and influencing factors existing between the different dynamic risk events. By dynamic risk factor identification, key factors with significant influence can be extracted from complex dynamic risk cluster data. Through Monte Carlo simulation, dynamic risk factor data can be randomly sampled and simulated for multiple times, so that results and influences occurring under different risk propagation situations are explored. By extracting the risk propagation time sequence data, the evolution process and the development trend of the risk event can be analyzed, and key time nodes and stages are identified. By extracting trend features of the risk propagation time sequence data, regularity and trending of the risk event development can be identified. By utilizing the risk trend characteristic data to conduct prediction processing, future risk development trends can be predicted and analyzed, potential risks can be found timely, and corresponding measures can be taken for coping.

Preferably, the weekly monitoring module comprises the following functions:

performing management and control quality calculation on risk trend prediction data through preset risk trend management and control indexes to obtain hidden danger point management and control quality data;

According to the hidden danger point control quality data, performing risk control response adjustment, and feeding back to a daily control module to perform control task response optimization processing to obtain optimized risk control task data;

Daily management and control quality warning processing is carried out on hidden danger point management and control quality data through optimizing risk management and control task data, so that management and control quality warning data are obtained;

And carrying out weekly control early warning processing on the responsible person through the risk control completion rate data and the control quality warning data to obtain weekly risk control early warning data.

The invention can identify the conditions of insufficient management and control or exceeding expectations by comparing with the preset management and control index. And the control response adjustment is carried out according to the hidden danger point control quality data, so that the control measures can be corrected in time, and the control effect and quality are improved. And the adjusted data is fed back to the daily management and control module for optimization treatment, so that the dynamic adjustment and optimization of the management and control task can be realized. The quality warning processing is carried out on the optimized management and control task data, the management and control quality problems or risk situations existing in early warning can be timely found out, the management and control early warning processing can be carried out on responsible persons every week through comprehensive analysis on the risk management and control completion rate data and the management and control quality warning data, the responsible persons are reminded of paying attention to the management and control problems and risk hidden dangers, corresponding measures are timely taken for coping and improving, and effective development and target achievement of risk management and control work are ensured.

Preferably, the item pre-warning treatment module comprises the following functions:

carrying out monthly early warning statistical processing on responsible persons through weekly risk management and control early warning data to obtain monthly early warning statistical data;

Deducting the personal performance according to the monthly early warning statistical data to obtain personal performance assessment data;

performing project performance punishment on weekly risk management and control early warning data through personal performance assessment data to obtain a risk management and control punishment strategy;

And after the construction of the dangerous large project is finished, carrying out risk marketing item processing based on a risk management and control punishment strategy to obtain project risk marketing item data.

According to the invention, monthly early warning statistical treatment is carried out on responsible persons through weekly risk management and control early warning data, quantitative evaluation and statistics are carried out on the performance of the responsible persons in the aspect of risk management and control, and management and control problems and risk hidden dangers are found out in time. And (3) deducting the personal performance according to the monthly early warning statistical data, and quantitatively evaluating and performing performance assessment according to the performance of responsible persons in the aspect of risk management and control, so as to perform corresponding performance punishment on responsible persons with poor performance. By comprehensively analyzing the personal performance assessment data, project performance penalties are carried out on responsible persons with poor performance, penalty measures including penalty, rewarding, saving and the like are carried out, so that the importance and execution force of the responsible persons on risk management and control work are enhanced. By carrying out risk marketing item processing according to a risk management and control punishment strategy, the case setting processing can be carried out on the risks which are solved or avoided, the project risks are cleared timely, and smooth completion of project construction and guarantee of project quality are ensured.

In this specification there is provided a computer readable storage medium having a computer program stored therein, wherein the computer program is arranged to execute the engineering project risk engineering integrated safety management system described in the above when run.

In this specification there is provided an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the engineering project risk engineering integrated safety management system described in the foregoing.

Drawings

FIG. 1 is a schematic diagram of an engineering project risk engineering integrated safety management system of the invention;

FIG. 2 is a functional flow diagram of the daily management and control module of FIG. 1;

FIG. 3 is a functional flow diagram of the risk supervision processing module of FIG. 1;

FIG. 4 is a flow chart of an implementation of an integrated safety management system for a large project risk;

The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.

Detailed Description

The following is a clear and complete description of the technical method of the present patent in conjunction with the accompanying drawings, and it is evident that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.

Furthermore, the drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus a repetitive description thereof will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. The functional entities may be implemented in software or in one or more hardware modules or integrated circuits or in different networks and/or processor methods and/or microcontroller methods.

It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.

In order to achieve the above objective, referring to fig. 1 to 4, the present invention provides an integrated safety management system for a critical engineering of an engineering project, comprising:

By acquiring the historical engineering case data, various potential risk factors and occurrence frequencies thereof can be identified, which is helpful for project teams to know risks in the engineering project planning stage, and preparation of precaution and control measures is carried out in advance, so that the possibility of risk occurrence is reduced. By analyzing the historical engineering case data, certain rules and trends can be found, and a reliable risk prediction model can be built. Such a model can predict the type and probability of risk that occurs by comparing and analyzing the conditions of the current engineering project. Based on the established risk prediction model, the specific situation of the current engineering project can be predicted, and corresponding risk data can be obtained. By evaluating and analyzing the engineering potential risk data, the risk scale of the engineering project, namely the degree of loss and the influence range can be determined. By constructing a multidimensional risk index system according to engineering potential risk data, risk factors can be systematically classified and generalized, so that comprehensive multidimensional risk assessment indexes are formed. By combining the risk scale data of the dangerous large project with the multidimensional risk index system, the risk of the project can be comprehensively evaluated. Through the risk matrix evaluation processing, the influence degree and possibility of different risk factors can be quantified, and corresponding risk evaluation grade data can be obtained. Based on the risk assessment grade data, risk items with higher influence degree and higher possibility can be determined, and corresponding counter measures and prevention lists are further determined. By performing daily risk management and control task dispatch on the effective prevention list data, project teams are ensured to know the risk items needing to be focused on the day in time, and the task dispatch can enable project managers and field staff to clearly know the daily risk tasks needing to be processed. Based on daily risk management and control task data, project teams can monitor and process real-time hidden dangers on site, and various potential hidden dangers and risks existing in engineering projects can be timely found through such real-time monitoring. The real-time hidden danger monitoring data and daily risk management and control task data are compared and analyzed, so that the completion condition of risk management and control work can be estimated. Hidden trouble context fusion processing is carried out on real-time hidden trouble monitoring data, and hidden trouble data of different time points and different monitoring dimensions can be integrated and fused. And dividing hidden danger data according to the similarity, and identifying hidden danger points of different categories and aggregation trends of the hidden danger points. By analyzing and mining the dynamic risk cluster data, the development trend and the evolution path of hidden danger points can be found. The effectiveness of risk management and control work can be evaluated and monitored by carrying out management and control quality calculation on risk trend prediction data. Based on the hidden danger point management and control quality data, the performance of the responsible person in the risk management and control work can be evaluated and monitored. A set of scientific and reasonable risk management and control punishment strategies can be established by carrying out performance punishment on responsible persons through weekly risk management and control early warning data. The strategy can evaluate the performance and performance of the responsible person in the risk management and control work, and proper punishment measures are adopted for the responsible person who fails to effectively fulfill the responsibilities, so that the responsible person is stimulated to participate in the risk management and control work more seriously, and the work quality and efficiency are improved. And according to the established risk management and control punishment strategy, marketing items are processed on the risks which are solved by taking effective measures. The processing can ensure the timely follow-up and implementation of risk management and control work, reduce accumulation and retention of potential risks and guarantee the smooth proceeding of projects. Therefore, the integrated safety management system for the engineering project and the dangerous large engineering carries out potential risk prediction on the dangerous large engineering project through three dimensions of advance, in-process and after-process, and establishes a risk relief strategy through a multi-dimensional risk index system. Daily risk management and control task processing is carried out according to a risk relief strategy, responsible person warning processing is carried out by considering weekly monitoring data, corresponding performance punishment is carried out according to warning information, the management responsibility of the floor is enhanced through quantitative assessment indexes and index completion and performance assessment hooks, and the safety risk of a danger-controlling large project in the construction process is effectively reduced.

In the embodiment of the present invention, as described with reference to fig. 1, a schematic diagram of an integrated security management system for a large-risk engineering of an engineering project according to the present invention is provided, where in the embodiment, the integrated security management system for a large-risk engineering of an engineering project includes:

In the embodiment of the invention, historical engineering case data is collected by utilizing various channels, and an engineering risk prediction model is established. The model is trained using project features and historical risk tags, including project location, scale, budget, and the like. Meanwhile, information such as keywords, phrases and structural information is extracted from the engineering project document to enrich the feature data. The engineering projects are classified into different types, risk assessment is facilitated, and factors such as characteristics, scale, geographic position and the like are considered for classification. And predicting the target engineering characteristic data by using the trained model, and outputting probabilities of different types of risks, such as financial, technical and supply chain risks. And further carrying out scale evaluation on the engineering potential risk data, and determining the influence degree on the project.

In the embodiment of the invention, a multi-dimensional risk index system is established by utilizing engineering potential risk data. And determining the weight of each risk factor through expert evaluation, and constructing a risk evaluation matrix by combining risk scale data of the risk major engineering. The matrix multiplies the risk factors by the weights and gathers to obtain an overall risk assessment score. The items are classified into different risk classes according to the scores. And examining and approving the effective prevention list according to the risk assessment grade, and obtaining feasible effective prevention list data through project management team examination and cost benefit assessment.

In the embodiment of the invention, a 'preventive measure list' data table is established to store effective measure information, wherein the effective measure information comprises fields such as measure names, execution time, execution personnel and the like. And setting an automatic dispatcher to perform a daily search list, generating a risk management task and sending a notification. And the executive checks the task and performs risk investigation on site, and simultaneously starts to monitor data acquisition in real time. Submitting the processing record to a daily management and control module, updating task data and auditing. And carrying out management and control completion rate analysis according to the task data and the real-time monitoring data to obtain risk management and control completion rate data.

In the embodiment of the invention, the image data is acquired from the real-time monitoring system, and the image quality is improved through the processes of a denoising filter, contrast adjustment and the like. And identifying hidden danger source objects and monitoring external scenes by utilizing an edge detection algorithm and a region segmentation technology. And fusing the environment information and hidden danger characteristics to obtain hidden danger monitoring fusion data. Preprocessing the fusion data and establishing a logistic regression model to obtain a real-time risk management and control level evaluation model. And (3) carrying out dynamic risk clustering calculation by using a K-means clustering algorithm, and understanding clustering characteristics. And finally, carrying out risk trend prediction by using a statistical method to obtain future risk trend prediction data.

In the embodiment of the invention, risk trend management and control indexes are set, and the deviation between predicted data and actual observed data is required to be not more than 10%. The deviation of a certain day is found to be more than 10%, and the mark management and control quality is unqualified. And adjusting a management and control strategy according to the management and control quality data, and increasing the inspection frequency and the supervision degree. And feeding back the regulated strategy to a daily management and control module to optimize the management and control task. Re-analyzing the quality data and finding that there is still a problem. Aiming at the management personnel to send out the management and control early warning notice, the management and control force is strengthened by reminding. And in consideration of low management and control task completion rate and a plurality of quality warnings, weekly management and control early warning notices are sent to responsible persons, so that management and control risk reduction is promoted.

In the embodiment of the invention, the performance of the responsible person is deducted by using weekly risk management and control early warning data, and a deduction standard is determined according to the early warning times and the severity of the quality problem. And penalty measures such as degradation, post adjustment, prize deduction and the like are adopted for responsible persons with poor performance. And according to the performance punishment result, a risk management and control punishment strategy of the project department is formulated, and punishment measures of different performance levels are defined. And summarizing and evaluating responsible persons according to the penalty strategy aiming at the finished dangerous engineering. And (3) realizing risk marketing item processing through a risk management and control punishment strategy, and completing the risk marketing item of the item.

In the embodiment of the invention, historical engineering case data and engineering project document data are acquired. Such data may come from a variety of sources, such as corporate internal databases, published engineering project databases, academic documents, and the like. A machine learning model is constructed to predict the risk encountered by the engineering project. For example, using supervised learning algorithms, such as decision trees, random forests, or neural networks, the model is trained from project features and risk labels in the historical engineering case data. For example, features of the item such as location, scale, budget, etc. are used as inputs to the model, and the risk actually encountered by the item is used as an output label. Useful information is extracted from the engineering project document to obtain more engineering feature data. This may include keywords, phrases, entities identified from the text, structural information of the document, and so on. Engineering projects are classified into different categories in order to better understand and evaluate their potential risks. This may be categorized based on factors such as the characteristics, scale, geographic location, etc. of the item. And predicting the target engineering characteristic data by using the trained engineering risk prediction model so as to evaluate the risk faced by the target engineering characteristic data. The model will output the probability that the item is exposed to different types of risks, such as financial risks, technical risks, supply chain risks, etc. Different types of risk scale evaluations are performed on engineering risk potential data to determine the extent of its impact on the project, such as the project risk impact scale.

In embodiments of the present invention, for example, there is a historical engineering case data set that includes various types of data, such as project names, locales, scales, budgets, durations, etc. In the preprocessing stage, the data are cleaned, repeated items are deleted, missing values are processed, abnormal values are removed and the like, the case that the project scale and the budget information are deleted is deleted, or obviously wrong data are corrected according to the business rules. Performing operations such as feature scaling, normalization, single-hot encoding, etc. on various features such as project size, budget, geographical location, etc., and creating new features such as project duration to budget ratio, cluster encoding of geographical location, etc. By recursive feature elimination, features that contribute less to the model performance are gradually removed by the performance of the model. For example, using a recursive feature elimination algorithm, in combination with cross-validation, non-significant features are progressively culled during the training of the model to improve the performance of the model. These features are classified by type, such as project-size features, geographic location features, budget features, etc., organized into a structured dataset. And (3) utilizing a dangerous engineering feature library, and finding out features closely related to engineering risks through data mining technologies (such as association rule mining, cluster analysis and the like). For example, it has been found that in certain geographical locations, the larger the engineering scale, the higher the engineering risk, or in certain projects, the supply of material may affect the progress and quality of the engineering. Training the risk factor characteristic data by using a preset convolutional neural network model, and adjusting parameters of the model by transfer learning so that the model can be better suitable for the characteristic data of a critical engineering characteristic library. During training, feature weights may also be optimized using techniques such as back-propagation algorithms to ensure that the model is better able to understand and utilize the risk factor feature data.

In the embodiment of the invention, a multidimensional risk index system is constructed according to engineering potential risk data and target engineering characteristic data so as to comprehensively evaluate various risks facing an engineering project. These risk factors are scored or ranked by expert discussions or expert evaluations to determine their relative importance and weight, and a risk assessment matrix is constructed in combination with the risk scale data of the critical large project. In this matrix, each risk factor is multiplied by its weight and then summed to yield the overall risk assessment score for the item. The items are classified into different risk classes, such as low risk, medium risk, high risk, etc., according to the size of the score. And according to the risk assessment grade data, a corresponding risk relief strategy is formulated so as to reduce the risk faced by the engineering project. These policies may include taking precautions, reinforcing supervision, adjusting project plans, and the like. The formulated risk preventive measure list is approved and it is determined which measures are effective and worthy of implementation. This can be done by a project management team review and discussion, or by an assessment based on factors such as the feasibility, cost effectiveness, etc. of each measure, resulting in effective prevention inventory data.

In the embodiment of the invention, historical risk occurrence frequency statistics is carried out on engineering potential risk data. This includes counting the number of occurrences of different types of risk in historical engineering cases, from which the probability of recurrence of each risk, i.e. the probability of reoccurrence in future engineering projects, can be calculated. The target engineering characteristic data is utilized to calculate the site management level index of the constructor, which can be completed by evaluating whether the constructor meets the safety production standardization requirement. These criteria may include safety facilities at the job site, staff training conditions, safety management regimes, etc. to evaluate the job site management level of the constructor, giving a corresponding safety production standardized review score. For example, a rating value is derived from the security production standardized review score. The safe production standardized evaluation score is carried out by adopting on-site actual score conversion. Safety production standardized review score = field actual score/(600-field section actual no term score involved) ×1000. From the engineering risk potential data, the emergency withstand capability of the target engineering, i.e. the capability of the project party to effectively cope when faced with an emergency risk event, can be assessed. This can be used to derive an emergency bearing capacity index for the project by evaluating the completeness and emergency handling capacity of the project side emergency plan. And carrying out cluster analysis on the historical engineering case data, and finding out different types of risk events including geological disasters, human accidents, supply chain interruption and the like. Classifying the risk events, such as classifying the geological disaster events into natural disaster types and classifying the human accident events into misoperation types, so as to obtain clustered historical risk event data. For example, in clustered historical risk event data, geologic hazard events are found to cause multiple casualties, while supply chain disruption events, while affecting project progress, do not cause casualties. The casualties of different types of risk events can be given according to the situations, such as the geological disaster event is a major casualties, and the supply chain interruption event is a general accident level. The economic loss level for each event is assessed by analyzing the direct loss (e.g., materials loss, equipment damage, etc.) and indirect loss (e.g., revenue loss due to production disruption, market reputation loss, etc.) caused by each event. The social impact loss is evaluated by analyzing the influence degree of each event on sensitive targets such as surrounding ecological environment, community safety and the like. Meanwhile, the social concern of each event can be mined through information such as media reports, public concern and the like so as to determine social influence loss data of the event. The fault impact level data is evaluated by analyzing the impact level of each event on infrastructure (such as equipment, power supply facilities, etc.), and the interruption level of life guarantee (such as water supply, power supply, communication, etc.) of the project. For example, the historical risk occurrence probability is 10%, the construction side field management level index is 80 minutes, the project emergency bearing capacity index is 70 minutes, and the risk possibility index is obtained by performing weighted average calculation based on the indexes. For example, casualties are rated as severe casualties, economic losses are rated as high economic losses, social impact losses are rated as medium social impact, and fault impact is rated as severe fault. From these data, a weighted average calculation may be performed to obtain a risk outcome severity indicator.

Preferably, the daily management module comprises the following functions:

As an example of the present invention, referring to fig. 2, a functional flow diagram of the daily management module in fig. 1 is shown, where the functions of the daily management module in this example include:

S201: the effective prevention list data are imported into a daily management and control module, and daily risk management and control task assignment is carried out on the effective prevention list data to obtain daily risk management and control task data;

In the embodiment of the invention, in the daily management and control module, a database or a data table is firstly required to be established and used for storing effective prevention list data. Such data may include details of precautions, execution time, executives, etc. Then, the effective prevention list data is imported into the database or data table. A data table named "preventive measure list" is created, including fields such as "measure name", "execution time", "executor", etc. Then, the preventive measure list data is imported into the data table in the format. By setting an automatic task scheduling program, effective prevention list data is automatically searched every day, and a daily risk management and control task is generated according to the execution time and the information of the execution personnel. The system can automatically distribute tasks and send notification to remind the executive personnel according to the emergency degree of the tasks, the workload of the executive personnel and other factors. For example, on a certain day the system automatically discovers the precautions that equipment inspection is needed, the system automatically assigns the task to the staff responsible for the equipment inspection and sends a task reminder.

S202: performing risk hidden danger check point processing according to daily risk management and control task data to obtain risk monitoring point data;

In the embodiment of the invention, an executive checks the risk management and control task distributed on the same day in the daily management and control module. The task content comprises risk hidden trouble points to be checked.

S203: performing on-site hidden danger real-time monitoring processing based on the risk monitoring point data to obtain real-time hidden danger monitoring data; a step of

In the embodiment of the invention, according to the task content, the operator performs on-site investigation on the site of the engineering project to check the existing risk hidden trouble points. And according to the risk monitoring point data, starting to acquire field data including images, videos, sensor data and the like in real time by using monitoring equipment. This involves equipment inspection, security sign inspection, work environment inspection, etc.

S204: transmitting the real-time hidden danger monitoring data to a daily management and control module for reporting the management and control task to obtain management and control task reporting data;

in the embodiment of the invention, after the executive completes the task, the record of the processing process is submitted to the daily management and control module. The data submission can be performed by form filling, file uploading and the like in the system. And transmitting the collected real-time monitoring data to a daily management and control module through a network. The transmission mode can be remote transmission based on the Internet or data transmission in a local area network. And after receiving the real-time monitoring data, the daily management and control module stores the real-time monitoring data in a corresponding database or data table. The data storage structure should be able to accommodate various dimensional information of the real-time monitoring data, such as time, location, monitoring values, etc. And after receiving the processing records submitted by the executive, the daily management and control module updates management and control task data and carries out auditing. The auditing process includes checking the validity and integrity of the submitted data.

S205: and carrying out management and control completion rate calculation on management and control task report data through daily risk management and control task data based on the daily management and control module to obtain risk management and control completion rate data.

In the embodiment of the invention, a daily management and control module acquires management and control task report data and daily risk management and control task data from a database. The management and control task report data includes a record of the treatment submitted by the daily executive and the daily risk management and control task data includes daily generated management and control task information. And for the successfully matched management and control task, searching whether a corresponding processing record exists in the management and control task report data. If the processing record exists, judging that the management and control task is finished; if no processing record exists, it is determined that the management and control task is not complete. For example, after the on-site manager initiates the task, the on-site manager is responsible for completing the report of the construction process record, the master worker completes the report of the technical review record, the safety master supervision completes the report of the hidden trouble investigation record, and the project manager completes the report of the lead zone record. And automatically generating the work completion rate of the day after the management and control record of the dangerous large project of the day is completed, wherein the completion rate is equal to the number of uploaded monitoring records of the dangerous large project initiated on the day/=the number of monitoring records which should be uploaded of the dangerous large project initiated on the day.

As an example of the present invention, referring to fig. 3, a functional flow diagram of the risk supervision processing module in fig. 1 is shown, where the functions of the risk supervision processing module in this example include:

S301: performing image frame enhancement processing on the real-time hidden danger monitoring data to obtain hidden danger monitoring image frame data;

In embodiments of the present invention, for example, a frame of image data is acquired from a real-time hazard monitoring system, which is typically acquired by a monitoring camera or sensor. Noise in the image is removed using a denoising filter (e.g., a gaussian filter or a median filter) to improve image quality. The contrast of the image is adjusted so that the details in the image are clearer and more prominent. A sharpening filter or edge enhancement algorithm is applied to enhance the edges and details of the image. Correcting the color deviation of the image to make the image color more accurate and natural.

S302: performing image segmentation processing according to the hidden danger monitoring image frame data to obtain hidden danger monitoring image segmentation data;

In the embodiment of the invention, the color image is converted into the gray image, the edge information in the image is detected by using an edge detection algorithm (such as Sobel and Canny algorithm), and the contrast of the image is adjusted, so that the target object in the image is more obvious. The image pixels are classified according to a preset threshold value to form a target area and a background area. Starting from the seed pixels, successive target regions are formed by step growth according to the similarity between pixels. The image is divided into regions according to the edge detection result, and the gap between the edges is regarded as a dividing line. The image is considered as a graph of pixels, and is partitioned into sub-graphs with similar properties by a graph theory algorithm (e.g., a min-cut max-flow algorithm). And marking the connected areas of the segmented image, and marking each connected area as a single object.

S303: carrying out hidden danger source object identification on hidden danger monitoring image segmentation data to respectively obtain hidden danger source object characteristic data and monitoring external scene data;

in the embodiment of the invention, for example, for each target area, the characteristics of the shape, the size, the color and the like of the target area are extracted first, and then whether the target area is a hidden danger source object is judged according to each identified object. If a region is identified as a crack, it is marked as a hidden trouble source object; if a region is not identified as a hidden trouble source object, it is marked as monitoring an external scene. At the same time, for each region in the monitored external scene, its characteristic data such as illumination intensity, surrounding environment, etc. are also extracted. And finally, respectively storing the characteristic data of the identified hidden danger source object and the monitored external scene data.

S304: hidden trouble context fusion processing is carried out on hidden trouble source object feature data through monitoring external scene data, so that hidden trouble monitoring fusion data are obtained;

In the embodiment of the invention, for example, a group of monitoring external scene data comprises information such as illumination intensity, weather condition and the like, and a group of hidden danger source object characteristic data comprises characteristics such as the size, the shape and the like of cracks. First, the two sets of data are region matched, ensuring that they correspond spatially. And then fusing the environment information in the monitored external scene data with the characteristic information of the hidden danger source object in a weighted average mode for the matched region to obtain fused data. For example, a comprehensive evaluation index can be obtained according to the weighted average of the illumination intensity and the size of the crack, and the comprehensive evaluation index is used for representing the hidden danger degree of the area. And finally, storing the fused data as hidden danger monitoring fusion data.

S305: performing risk management and control level assessment according to hidden danger monitoring fusion data to obtain real-time risk management and control level data; carrying out dynamic risk clustering processing according to the hidden danger monitoring fusion data to obtain dynamic risk clustering data;

In the embodiment of the invention, the type, the size and the environmental factors of the hidden danger are selected from the hidden danger monitoring fusion data as the characteristics for evaluating the risk. First, the features are subjected to data preprocessing, including processing of missing values and outliers, and data normalization. Then, a logistic regression model is built as a risk assessment model. And then, inputting the preprocessed data into a model for training, and establishing a real-time risk management and control level assessment model. And finally, inputting the real-time hidden danger monitoring fusion data into an evaluation model to obtain real-time risk management and control level data at each moment. A K-means clustering algorithm is selected as the clustering model. And then, inputting hidden danger monitoring fusion data into a K-means clustering algorithm to perform clustering calculation, and obtaining different dynamic risk clusters. And then, analyzing the clustering result to understand the dynamic risk characteristics represented by each cluster. And finally, storing the calculated dynamic risk clustering data.

S306: and carrying out risk trend prediction processing on the real-time risk management and control level data through the dynamic risk cluster data to obtain risk trend prediction data.

In the embodiment of the invention, features, such as time stamps, risk levels and the like, are extracted from dynamic risk cluster data. Real-time risk management level data is collected, including risk scores, number of risk events, etc. And carrying out time serialization processing on the dynamic risk cluster data so as to carry out statistical analysis. And (3) cleaning the real-time risk management and control level data, including removing abnormal values, filling missing values and the like. And selecting a proper statistical method for risk trend prediction, such as a moving average method, an exponential smoothing method or an ARIMA model, and the like. And carrying out time sequence model parameter estimation on the real-time risk management level data, such as smoothing coefficients in an exponential smoothing method. And predicting the latest dynamic risk cluster data by using the trained model and parameters to obtain risk trend prediction data in a future period of time.

In the embodiment of the invention, an association rule mining algorithm (such as an Apriori algorithm or an FP-Growth algorithm) is used for finding the association relation among the items in the data set. Appropriate support and confidence thresholds are set to filter out frequent item sets and strong association rules. For example, the dynamic risk cluster data contains attribute information of different risk events, such as location, time, risk type, etc. Factors having an important impact on risk variation, such as location, time, weather, etc., are identified using statistical methods and analysis of the data, these key risk factors are determined, and relevant features are extracted from the data, using dynamic risk factor data as input, including historical variation data for the risk factors. For example, the probability distribution type, such as normal distribution or uniform distribution, of each risk factor is determined, and the corresponding parameters are set. Each risk factor is randomly sampled, generating a large number of random samples. And finally, performing Monte Carlo simulation by using the random samples and model parameters to obtain a scene of dynamic risk propagation. And carrying out time sequence analysis on each piece of dynamic risk propagation simulation data, and extracting time sequence information of risk propagation, including time points, risk degrees and the like. And carrying out statistical analysis on the risk propagation time sequence data, and extracting various trend characteristics such as slope, change rate, periodicity and the like of the trend. For example, statistical analysis of the data extracts various trend features such as average rate of change of risk level, maximum, minimum, fluctuation amplitude, etc. Meanwhile, the data is smoothed by using a time sequence analysis method, such as a moving average method, so that the periodicity and the variation trend of the trend are further extracted. The ARIMA model was selected for risk trend prediction. First, the ARIMA model is trained using historical data to obtain model parameters. And then, inputting the real-time risk management and control level data into a model for prediction, and obtaining risk trend prediction data in a future period of time.

Preferably, the weekly monitoring module comprises the following functions:

In the embodiment of the present invention, for example, the preset risk trend control index is that the deviation between the risk trend prediction data and the actual observation data should be within 10% in a specific time period. The risk trend prediction data is compared with the actual observation data, and the prediction deviation of a certain day is found to be more than 10%. And obtaining that the management and control quality index of the day is unqualified through statistical calculation. According to the control quality warning data obtained in the previous step, the control quality of some hidden danger points is found to be problematic, mainly due to inaccuracy of some risk trend predictions. And the risk management and control strategies of the hidden danger points are adjusted, so that the inspection frequency and the supervision degree are increased. And then, feeding back the adjusted control strategy to the daily control module, and adjusting and optimizing the corresponding control task so as to improve the control quality. And re-analyzing the management and control quality data of hidden danger points according to the optimized risk management and control task data, and finding that some management and control quality problems still exist. For example, the control quality of a certain hidden trouble point is always above a warning line, and needs to be processed in time. And sending a control quality warning notice to relevant responsible persons, and requiring the relevant responsible persons to strengthen the supervision and improve the control measures so as to ensure the safety of hidden danger points. And comprehensively considering the risk management and control completion rate data and the management and control quality warning data, finding that the management and control task of a certain responsible person is lower in completion rate, and receiving a plurality of hidden danger point management and control quality warnings. Therefore, the weekly control early warning notice is sent to the responsible person, the responsible person is reminded to strengthen the supervision force on hidden danger points, and effective measures are taken to reduce the control risk.

In the embodiment of the invention, weekly risk management and control early warning data are used as input. And carrying out statistical calculation on the risk management and control early warning data received by each responsible person in each month. The statistical calculation comprises the early warning times received by each responsible person, the number of hidden danger points involved, the severity of the quality control problem and the like. And correspondingly deducting the personal performance of each responsible person according to the monthly early warning statistical data. The deduction standard can be determined according to the early warning times, the number of hidden danger points involved, the severity of the control quality problem and other factors. For example, responsible person a receives 5 risk management warnings within a month, because 2 of them belong to a serious quality problem, each serious quality problem warning will be deducted by 5% of the individual performance according to the corporate specifications, so that responsible person a's performance will be deducted by 10% in that month. And according to the personal performance assessment data, summarizing and evaluating the performance of each responsible person in the project department within a certain period of time. For responsible persons with poor performance, corresponding penalty measures are taken, such as degradation, tuning, prize deduction, etc. And according to the performance penalty result, a risk management and control penalty strategy of the project department is formulated, and penalty measures to be suffered by responsible persons with different performance levels are defined. And summarizing and evaluating the punishment conditions of responsible persons according to the risk management and control punishment strategy aiming at the finished dangerous engineering. For example, when the construction of the dangerous large project is finished, a field manager initiates a finishing application, a security master audit and a project approval and then automatically finishes the project approval, and the project approval is fed back to a system home page without daily monitoring, which belongs to a risk sales item.

The integrated safety management system for the engineering project and the dangerous engineering has the beneficial effects that the integrated safety management system for the engineering project and the dangerous engineering carries out potential risk prediction and management on the dangerous engineering project through three dimensions of advance, in-process and after-process. Firstly, through analysis of historical engineering case data, potential risk factors and occurrence frequencies thereof are identified, and a reliable risk prediction model is established. And secondly, predicting the risk of the current engineering project based on a risk prediction model to form corresponding risk data. And then, constructing a multidimensional risk index system, and carrying out system classification and induction on risk factors to carry out comprehensive evaluation and treatment. And quantifying the influence degree and possibility of different risk factors through risk matrix evaluation, and determining a countermeasure and prevention list. Daily risk management and control task dispatching ensures that project teams timely know risk items needing important attention, and real-time hidden danger monitoring discovers potential hidden danger and risks existing in engineering projects. And then integrating and fusing the real-time hidden danger monitoring data, identifying hidden danger points and aggregation trends thereof, and predicting the development trend and the evolution path of the hidden danger points. And finally, performing performance punishment on responsible persons according to weekly risk management and control early warning data, establishing a scientific and reasonable risk management and control punishment strategy, promoting the responsible persons to participate in the risk management and control work more seriously, and reducing the safety risk of engineering projects. All modules have system relevance, and the system has integrity, so that the system can better help work projects to strengthen dangerous engineering management, and can realize responsibility people. The management responsibility of the floor is enhanced by quantifying the assessment indexes, finishing the indexes and hooking the performance assessment. Through the list of the home page system, the company can better grasp the large project management condition Shi Wei.

The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The integrated safety management system for the engineering project and the dangerous engineering is characterized by comprising the following modules:

2. The engineering project risk major engineering integrated safety management system according to claim 1, wherein the risk identification evaluation module includes the following functions:

3. The engineering project risk large engineering integrated safety management system according to claim 2, wherein the engineering risk prediction model training according to the historical engineering case data is performed, and comprises the following steps:

4. The engineering project risk major engineering integrated safety management system according to claim 2, wherein the multi-dimensional risk index construction module includes the following functions:

5. The integrated safety management system for engineering project risk and large engineering according to claim 4, wherein the multidimensional risk index system includes a risk likelihood index and a risk consequence severity index, and the steps of constructing the multidimensional risk index system according to engineering potential risk data and target engineering characteristic data include:

6. The engineering project risk major engineering integrated safety management system according to claim 1, wherein the daily management and control module includes the following functions:

7. The engineering project risk major engineering integrated safety management system according to claim 1, wherein the risk supervision processing module includes the following functions:

8. The integrated safety management system for engineering project risk major engineering according to claim 7, wherein when executing the risk trend prediction processing on the real-time risk management level data by the dynamic risk cluster data, the system comprises the following steps:

9. The engineering project risk major engineering integrated safety management system according to claim 1, wherein the weekly monitoring module includes the following functions:

10. The engineering project risk large engineering integrated safety management system according to claim 1, wherein the project pre-warning treatment module comprises the following functions: