Flammable and explosive process production line risk assessment method, device, equipment and medium

Abstract

The invention discloses a flammable and explosive process production line risk assessment method, device, equipment and medium, and relates to the field of flammable and explosive process production lines. The method comprises the following steps: carrying out risk analysis by utilizing a danger and operability analysis method to obtain risk factors, collecting subjective data and objective data from five dimensions of a man-machine material method for each risk factor, and constructing a five-dimensional data set; carrying out initial assignment on the risk factors by utilizing a Delphi method, carrying out weighting on an initial risk assignment result by utilizing a network analytic hierarchy process, and carrying out space-time slice decomposition on the risk sections to obtain space-time fragment data containing weight data; constructing a dynamic cloud model based on the spatio-temporal fragment data, generating a standard cloud atlas and a current dynamic cloud atlas based on historical normal working condition data and current working condition data, and performing offset degree analysis on the cloud atlas to obtain a risk assessment result; the objective consistency and dynamic space-time correlation analysis capability of risk assessment can be effectively improved, the accuracy is high, and the assessment efficiency is high.

Description

Technical Field

[0001] The present invention belongs to the technical field of flammable and explosive process production lines, and in particular relates to a risk assessment method, device, equipment and medium for flammable and explosive process production lines. Background Art

[0002] Flammable and explosive process production lines usually involve energetic materials and toxic and hazardous substances, with a high level of basic risk. In addition, there are many manual steps and a large number of people involved. The entire production line has a high hazard level and a low degree of quantification.

[0003] Currently, safety risk identification and qualitative assessments for flammable and explosive production lines primarily rely on a hazard and operability analysis (HRA) method based on expert judgment. This phased assessment method is not only highly subjective and lacks quantitative assessment tools, but also prone to significant fluctuations in safety risk assessments when the number of invited experts changes or is insufficient, leading to inaccurate results. Furthermore, reliance on expert judgment results in a long and cyclical risk analysis cycle, lacking dynamism and continuity, making it difficult to accurately control risk levels during production. Summary of the Invention

[0004] In order to solve the above problems, the present invention proposes a risk assessment method, device, equipment and medium for flammable and explosive process production lines, which effectively improve the objective consistency and dynamic spatiotemporal correlation analysis capabilities of risk assessment, have high accuracy and high assessment efficiency, and are suitable for large-scale promotion and use.

[0005] In a first aspect, the present invention provides a risk assessment method for a flammable and explosive process production line, comprising:

[0006] S1, obtaining risk factors for multiple production sections of a flammable and explosive process production line; wherein the risk factors are obtained by performing a risk analysis on the multiple production sections of the flammable and explosive process production line using a hazard and operability analysis method;

[0007] S2, for each of the risk factors, collect subjective and objective data from five dimensions: personnel dimension, machine and equipment dimension, raw material dimension, process method dimension, and environment dimension to construct a five-dimensional data set;

[0008] S3, based on the five-dimensional data set, the Delphi method is used to perform initial assignment of risk factors to obtain the initial risk assignment results;

[0009] S4, based on the mutual dependence and feedback relationship between the risk factors, using the network analysis hierarchy process to weight the initial risk assignment results to obtain weight data;

[0010] S5, performing spatiotemporal slicing decomposition on the production section with risk points of the flammable and explosive process production line according to the time axis to obtain a plurality of spatiotemporal segment data; wherein each of the spatiotemporal segment data includes the weight data within a specific time period;

[0011] S6, constructing a dynamic cloud model based on the spatiotemporal segment data and cloud model theory;

[0012] S7, based on the historical normal operating condition data of multiple production sections of the flammable and explosive process production line, using the aforementioned S1 to S5 to generate standard weight data, and generating a standard cloud map based on the standard weight data and the dynamic cloud model; and based on the current operating condition data of multiple production sections of the flammable and explosive process production line, using the aforementioned S1 to S5 to generate current weight data, and generating a current dynamic cloud map based on the current weight data and the dynamic cloud model;

[0013] S8, determining a deviation based on the standard cloud map and the current dynamic cloud map, and determining a risk assessment result according to the deviation.

[0014] In an optional embodiment, the subjective data is determined by expert evaluation, and the objective data is determined by sensor detection.

[0015] In an optional embodiment, the personnel dimension includes employee quality, operating specifications and safety responsibilities, the machine and equipment dimension includes equipment status, safety protection and equipment updates, the raw material dimension includes raw material quality and storage conditions, the process method dimension includes process flow and quality control, and the environment dimension includes production environment, safety facilities and environmental protection requirements; wherein, the personnel dimension, raw material dimension and process method dimension use expert evaluation to collect subjective data, and the machine and equipment dimension and environment dimension use sensor equipment to collect objective data.

[0016] In an optional embodiment, the S7 includes:

[0017] The standard cloud image and the dynamic cloud image are generated by using forward cloud transformation and backward cloud transformation algorithms.

[0018] In an optional embodiment, the S3 includes:

[0019] A scoring standard is set for the five-dimensional data set corresponding to each risk factor, and the initial assignment is determined in combination with expert opinions and the scoring standard.

[0020] In an optional embodiment, the S6 includes:

[0021] Generate a first normal random number Enn, the first normal random number Enn is a random number about the entropy En of the spatiotemporal segment data, the first normal random number Enn~N(En,He 2); wherein, He is the super entropy of the space-time segment data; generating a second normal random number x i , the second normal random number x i is the degree of membership, the second normal random number x i ~N(Ex,Enn 2 ), i is a positive integer; where Ex is the expectation of the spatiotemporal segment data; the membership degree μ(x i ):

[0022]

[0023] Repeat the previous step until the total number of cloud droplets is generated;

[0024] Based on the total cloud droplet number and the inverse cloud generation algorithm, cloud model parameters (Ex, En, He) are generated; wherein,

[0025]

[0026] In the above formulas (2), (3) and (4), m is a positive integer, S 2 Represents x i The variance of .

[0027] In an optional embodiment, after step S8, the method further includes:

[0028] Determining contribution data of the risk factors according to the risk assessment results;

[0029] Updating the weight data according to the contribution data;

[0030] The standard cloud map is updated according to the updated weight data.

[0031] In a second aspect, the present invention provides a risk assessment device for a flammable and explosive process production line, comprising:

[0032] A data acquisition module is used to obtain risk factors of multiple production sections of a flammable and explosive process production line; wherein the risk factors are obtained by performing a risk analysis on the multiple production sections of the flammable and explosive process production line using a hazard and operability analysis method;

[0033] A subjective and objective data determination module is used to collect subjective and objective data from five dimensions: personnel dimension, machine and equipment dimension, raw material dimension, process method dimension, and environment dimension for each risk factor, and construct a five-dimensional data set;

[0034] The assignment module is used to perform initial assignment of risk factors based on the five-dimensional data set using the Delphi method to obtain the initial risk assignment results;

[0035] A weighting module is used to weight the initial risk assignment result based on the mutual dependence and feedback relationship between risk factors using the network analysis method to obtain weight data;

[0036] A spatiotemporal decomposition module is used to perform spatiotemporal slicing decomposition of production sections with risk points in flammable and explosive process production lines according to the time axis to obtain a plurality of spatiotemporal segment data; wherein each of the spatiotemporal segment data includes the weight data within a specific time period;

[0037] A model building module, configured to build a dynamic cloud model based on the spatiotemporal segment data and cloud model theory;

[0038] a quantitative conversion module for generating standard weight data based on historical normal operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module, subjective and objective data determination module, value assignment module, and spatiotemporal decomposition module, and generating a standard cloud map based on the standard weight data and the dynamic cloud model; and, based on current operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module, subjective and objective data determination module, value assignment module, and spatiotemporal decomposition module, and generating a current dynamic cloud map based on the current weight data and the dynamic cloud model;

[0039] The risk assessment module is used to determine the deviation degree based on the standard cloud map and the current dynamic cloud map, and determine the risk assessment result according to the deviation degree.

[0040] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in any one of the aforementioned embodiments when executing the computer program.

[0041] In a fourth aspect, the present invention provides a computer-readable medium having a non-volatile program code executable by a processor, wherein the program code enables the processor to execute the method described in any one of the aforementioned embodiments.

[0042] The beneficial effects of the technical solution provided by the embodiment of the present invention are as follows: the risk assessment method, device, equipment and medium for flammable and explosive process production lines proposed by the present invention, because the risk factors are obtained by risk analysis of multiple production sections of the flammable and explosive process production line using the hazard and operability analysis method, and because the subjective and objective data are collected from five dimensions of personnel, machinery and equipment, raw materials, process methods and environment, and a five-dimensional data set is constructed, it can comprehensively cover the man-machine-material-method-environment risk factors of the flammable and explosive production line, break through the limitations of traditional single reliance on expert experience, and realize risk quantitative assessment combining subjective and objective factors, thereby improving the completeness and accuracy of risk factor coverage; due to the use of German The Alphi method initially assigns values ​​to risk factors and combines it with the network analysis method to weight the dependencies between factors, thereby organically integrating the objective data collected by sensors with the subjective opinions of experts. The dynamic correlation of risk factors is quantified through ANP, which solves the assessment bias problem caused by ignoring the feedback relationship between factors in traditional methods, improves the objectivity and logical consistency of risk assessment, and determines risks based on the deviation between the standard cloud map and the current dynamic cloud map. The staged assessment (standard cloud map) is combined with the real-time dynamic assessment (current cloud map). The cloud model's ability to dynamically represent multi-source heterogeneous data (subjective and objective data) is used to achieve continuous temporal and spatial evolution tracking of production line risks, significantly enhancing the early warning capability of high-risk working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A schematic diagram of a process flow for a risk assessment method for a flammable and explosive process production line provided by an embodiment of the present invention;

[0044] Figure 2 A schematic diagram illustrating the principle of analyzing risk factors from five dimensions provided by an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of the principle of bidirectional conversion of dynamic cloud models provided by an embodiment of the present invention;

[0046] Figure 4 A dynamic cloud map of production line risks provided by an embodiment of the present invention;

[0047] Figure 5 Another schematic flow chart of the risk assessment method for flammable and explosive process production lines provided by an embodiment of the present invention;

[0048] Figure 6 A schematic diagram of the system principle of the flammable and explosive process production line risk assessment device provided by an embodiment of the present invention;

[0049] Figure 7 A schematic diagram of the system principle of an electronic device provided by an embodiment of the present invention.

[0050] In the figure: 10-data acquisition module; 20-subjective and objective data determination module; 30-assignment module; 40-weighting module; 50-time-space decomposition module; 60-model construction module; 70-quantitative conversion module; 80-risk assessment module; 1000-electronic device; 1001-communication interface; 1002-processor; 1003-memory; 1004-bus. DETAILED DESCRIPTION

[0051] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0052] Reference Figure 1 A risk assessment method for a flammable and explosive process production line includes the following steps S1 to S8.

[0053] Step S1, obtaining risk factors of multiple production sections of a flammable and explosive process production line; wherein the risk factors are obtained by performing a risk analysis on multiple production sections of the flammable and explosive process production line using a hazard and operability analysis method.

[0054] Specifically, flammable and explosive process lines refer to industrial production lines that involve flammable substances (such as alcohol and hydrogen) or explosive substances (such as nitroglycerin and liquefied petroleum gas) during the production process. These substances can cause combustion or explosion under certain conditions (such as high temperature, high pressure, and static sparks). For example, in the metallurgical industry, when refining steel ingots, a mixed gas atmosphere such as argon and hydrogen is required to prevent metal oxidation. However, hydrogen has an extremely high risk of combustion and explosion, making this a typical flammable and explosive process line. Another example is the storage and transfer process in toluene tank areas involved in organic solvent production. Toluene is a highly flammable liquid (flash point 4.4°C, explosion limit 1.2% to 7.1%). Its vapor, when mixed with air, can cause explosions when exposed to static sparks or high temperatures, also a typical flammable and explosive process line. Other examples include the distillation section of gasoline production in chemical plants, the electrolyte injection process in lithium battery manufacturing, and the mixing and stirring process in gunpowder production. Once an accident occurs in these flammable and explosive processes, it is easy to lead to catastrophic consequences such as fire, explosion, and toxic leakage. Therefore, parameters such as temperature, pressure, and concentration need to be strictly monitored in daily production.

[0055] Risk factors refer to potential sources of danger or conditions that may lead to accidents.

[0056] The Hazard and Operability Study (HAZOP) is a systematic and structured risk assessment method that identifies potential hazards and proposes control measures by analyzing the deviations between process parameters and design intent. The HAZOP can specifically include the following steps (1) to (6).

[0057] (1) Decompose complex processes into independent units.

[0058] (2) Clarify the normal operating parameters of each node, for example, the temperature range is 20℃ to 30℃ and the pressure is ≤0.5MPa.

[0059] (3) Use standardized vocabulary to systematically infer bias, such as "none", "excess", "reverse", etc.

[0060] (4) Perform deviation analysis, for example, deviation analysis of excessive temperature.

[0061] (5) Rating the severity and likelihood of identified hazards.

[0062] (6) Formulate engineering improvements, operating procedures or emergency plans.

[0063] In some possible embodiments, a risk factor analysis is also performed using a process review method. After the risk factor analysis, risk factors for multiple risky sections are obtained (in some scenarios, not all sections will have risk factors), and each section may have one or more risk factors.

[0064] In step S2, for each risk factor, subjective and objective data are collected from five dimensions: personnel dimension, machine equipment dimension, raw material dimension, process method dimension, and environment dimension, to construct a five-dimensional data set.

[0065] Here, subjective data is determined through expert evaluation, while objective data is determined through sensor testing. Each risk factor is analyzed from five dimensions: personnel, machinery and equipment, raw materials, process methods, and environment. Expert evaluation (Delphi method) is used to collect subjective data for personnel, raw materials, and process methods, while sensor equipment is used to collect objective data for machinery and equipment and environment.

[0066] It should be noted that the environmental dimension sometimes also uses expert assessments to collect subjective data. This means that depending on the scenario, the environmental dimension may collect only objective data, only subjective data, or both. It should be emphasized that the use of subjective or objective data for the five dimensions above is generally considered, and this may be adjusted based on specific needs during implementation.

[0067] The five-dimensional risk analysis method described above quantitatively assesses the impact of five factors—people, machinery, materials, processes, and the environment—on production line safety. This provides companies with a scientific basis for decision-making, optimizing production safety management and reducing safety risks. The following is a detailed introduction to the five-dimensional analysis.

[0068] (1) Personnel dimension (people)

[0069] The personnel dimension includes indicators such as employee quality, operating standards, and safety responsibility. Employee quality is used to assess employee safety training, skill levels, and safety awareness, and data is collected through questionnaires and practical assessments. Operating standards are used to assess whether employees follow standard procedures, with the number of violations and accident rates serving as quantitative indicators. Safety responsibility clarifies the safety responsibilities of employees at all levels and is assessed through safety responsibility letters and accountability systems.

[0070] (2) Machine dimensions (machine)

[0071] The machine and equipment dimension includes indicators such as equipment status, safety protection, and equipment updates. Equipment status is used to assess equipment availability, failure rates, and maintenance status, and is quantified using data from the equipment management system. Safety protection verifies the equipment's safety precautions and emergency shutdown devices, obtaining data through on-site inspections and testing. Equipment updates typically consider the impact of equipment aging and technological advancements on safety, combined with factors such as equipment age and maintenance costs.

[0072] (3) Raw material dimensions (material)

[0073] The raw material dimension includes indicators such as raw material quality and storage conditions. Raw material quality is used to evaluate the purity, stability, and harmful substance content of raw materials, and is quantified through quality inspection reports. Storage conditions are used to examine the storage environment, temperature and humidity control, fire and explosion prevention measures, etc. of raw materials, and are subject to on-site inspection and record keeping.

[0074] (4) Process method dimension (method)

[0075] The process approach dimension includes indicators such as process flow and quality control. The process flow is used to evaluate the rationality and safety of the process and is reviewed through process flow charts, operating instructions, etc. Quality control is used to examine quality control measures in the production process, such as quality inspection and handling of non-conforming products, and is analyzed through quality reports.

[0076] (5) Environmental dimension (environment)

[0077] The environmental dimension includes indicators such as the production environment, safety facilities, and environmental protection requirements. The production environment assesses the physical environment of the workshop, including temperature, humidity, noise, and lighting, with data collected through environmental monitoring equipment. Safety facilities verify the completeness and effectiveness of firefighting facilities, emergency rescue equipment, and safe passages, acquiring data through on-site inspections and testing. Environmental protection requirements consider environmental protection measures implemented during the production process, such as wastewater, exhaust gas, and solid waste treatment, to ensure compliance with national environmental standards.

[0078] In this embodiment, the hazard and operability analysis method is used to conduct a phased basic risk analysis, explore the risk scenarios of key control points in the production process, analyze key risk points, and use digital means (such as installing sensors and other detection equipment) to collect risk point information and conduct quantitative risk assessment to construct a five-dimensional data set. In view of the shortcomings of low automation levels and small amounts of data collection in most flammable and explosive production lines, a qualitative and quantitative evaluation model for production line safety risks is established from the five perspectives of people, machines, materials, methods, and environment. Through a risk assessment method that combines subjective and objective factors, the risk level of the current production line is accurately analyzed, and the safety risks of the production line can be comprehensively and systematically evaluated, providing enterprises with a scientific basis for decision-making. Through regular evaluation, dynamic adjustment and continuous improvement, the level of production safety management of enterprises can be continuously improved and safety risks can be reduced.

[0079] The following examples further illustrate these five dimensions. The application of this method (model) is demonstrated using the recrystallization process of a specific plant. The recrystallization process includes six steps (i.e., the aforementioned production stages): raw material preparation, crystallization, aging, drying, and packaging. First, a basic risk analysis of each process step was conducted using the process review method and the hazard and operability analysis method. Details are shown in the table below.

[0080] Table 1 Key points for risk analysis of the process steps of the recrystallization production line

[0081]

[0082]

[0083]

[0084] Step S3: Based on the five-dimensional data set, the risk factors are initially assigned using the Delphi method to obtain the initial risk assignment results.

[0085] Specifically, scoring criteria are set for the five-dimensional data set corresponding to each risk factor, and the initial assignment is determined by combining expert opinions and scoring criteria.

[0086] During implementation, the Delphi method, also known as the expert opinion method, was used to initially assign values ​​to the five-dimensional data. This process leverages the wisdom and experience of industry experts, ensuring the authority and accuracy of the assessment. The Delphi method is a structured expert consensus forecasting and decision-making method that solicits expert opinions through multiple rounds of anonymous questionnaires and iteratively feeds back statistical results, gradually converging the expert group's judgment and ultimately forming a highly credible collective conclusion.

[0087] Among them, before assignment, the five-dimensional data set is used as input, and the data of the subjective dimension in the personnel dimension, machine equipment dimension, raw material dimension, process method dimension and environmental dimension are initially assigned using the Delphi method, and the objective data collected by the sensor is directly used to calculate the initial assignment (such as calculating the assignment through statistical methods).

[0088] Considering the low level of automation in current production lines and the lack of historical data records, this embodiment uses the Delphi method to assign initial risks. Through subsequent automated digital information transformation, relevant risk point information is collected. The risk assessment model is modified through a supervised learning network model, and objective data analysis results are introduced to achieve a quantitative analysis of production line safety risks that integrates subjective and objective factors.

[0089] Step S4: Based on the mutual dependence and feedback relationship between risk factors, the initial risk assignment results are weighted using the network analysis method to obtain weight data.

[0090] Usually, the corresponding weight is assigned to each factor according to its importance to production line safety. Here, the Analytic Network Process (ANP) is used to assign more precise weights to risk factors. For each process link, risk assessment is conducted from five dimensions: people (personnel), machines (machine equipment), materials (raw materials), methods (process methods), and environment (environment). A multi-level risk transfer path is constructed, analyzed, and a reliable risk transfer map and analysis model are established. Finally, the risk assessment results are output. For example, Figure 2 As shown, a dedicated network analytic hierarchy process tool (e.g., yaanp) is used to input the raw material preparation process, crystallization process, aging process, drying process, and packaging process with risk factors into the tool to establish a multi-level risk transmission path, such as Figure 2 The raw material preparation process analyzes risk factors from five dimensions: operating specifications (people), vibration (machine), impurity content (material), material ratio (method), and environmental dimension (environment).

[0091] The network analysis method takes into account the interdependence and feedback relationship between factors, making the risk assessment model closer to the actual production situation and improving the reliability of the analysis.

[0092] Step S5: Spatiotemporal slicing of the production sections of the flammable and explosive process line with risk points is performed along the time axis to generate multiple spatiotemporal slice data. Each spatiotemporal slice data includes weighted data for a specific time period. The specific time period is determined based on expert opinion and process characteristics.

[0093] This embodiment introduces the concept of space-time slicing, performs slice analysis on sections such as feeding and crystallization, and discretizes continuous space-time data into a data processing method for multi-dimensional fragments. Space-time slicing aims to achieve collaborative cutting through time dimension segmentation and space dimension partitioning to capture the evolution characteristics of dynamic systems. The time dimension can divide the process production line operation data into continuous or overlapping time windows, such as minute-level slices or hour-level slices, so as to capture the temporal evolution law of risk factors. The spatial dimension can divide spatial units according to the physical location or function of the production section and analyze the transmission path of risks in the spatial network. The time window and the spatial unit are cross-combined to form space-time fragment data (time-space-data) to quantify the dynamic characteristics of risks in the space-time coupling field.

[0094] In this embodiment, step S5 may include the following steps (1) to (3).

[0095] (1) Determine the time window length and spatial unit division rules (e.g., work section boundaries), divide the operation data (including weight data) into time slices according to the time window length, and classify them by spatial units within each slice.

[0096] (2) The spatiotemporal segments corresponding to the weight data in step S4 are converted into spatiotemporal slice data.

[0097] (3) Establish a unique identifier for the spatiotemporal segment (such as timestamp + section ID) to facilitate subsequent model calls and data backtracking.

[0098] This embodiment innovatively introduces the concept of time-space slicing. Targeting the key production stages of flammable and explosive materials, it decomposes the complex production process into a series of manageable segments through slice analysis on the timeline. Each segment contains operational data within a specific time period. The generated time-space segment data facilitates an in-depth understanding of the dynamic changes in the production process.

[0099] Step S6: constructing a dynamic cloud model based on the spatiotemporal segment data and cloud model theory, which specifically includes the following steps (1) to (5).

[0100] Step (1) generates a first normal random number Enn, which is a random number about the entropy En of the spatiotemporal segment data. The first normal random number Enn~N(En,He 2 )(That is, Enn has a mean of En and a variance of He 2 Normal distribution); where He is the super entropy of the spatiotemporal segment data;

[0101] Step (2), generate the second normal random number x i , the second normal random number x i is the membership degree, the second normal random number x i~N(Ex,Enn 2 )(i.e. x i The mean is Ex and the variance is Enn 2 Normal distribution), i is a positive integer; where Ex is the expectation of the spatiotemporal segment data;

[0102] Step (3), determine the membership degree by the following formula (1):

[0103]

[0104] Step (4), repeating steps (1) to (3) above until the total number of cloud droplets is generated;

[0105] Step (5) generates cloud model parameters (Ex, En, He) based on the total number of cloud droplets and the inverse cloud generation algorithm (BCT algorithm); where,

[0106]

[0107] In the above formulas (2), (3) and (4), m is a positive integer, S 2 Represents x i The variance of .

[0108] Production line operating status data acquired through sensors and expert experience is highly uncertain and fuzzy. The cloud model establishes a novel bidirectional conversion method between qualitative fuzzy concepts and quantitative, precise data. This method can simultaneously account for both fuzziness and randomness in evaluation and has been widely applied to uncertainty conversion problems. This method is a bidirectional cognitive model that combines probability theory and fuzzy set theory to address problems under uncertainty. Qualitative concepts in the CM are characterized by three characteristic parameters: expectation (Ex), entropy (En), and hyperentropy (He). Expectancy measures the fundamental certainty of a qualitative concept and is its most representative point. Entropy measures the uncertainty of a qualitative concept. On the one hand, entropy reflects the size of the cloud droplet cluster that can be accepted by linguistic values ​​in the number domain, i.e., the degree of fuzziness. On the other hand, entropy also reflects the randomness of the cloud droplets representing the qualitative concept. Hyperentropy measures the uncertainty of entropy, i.e., the entropy of entropy.

[0109] Step S7, based on the historical normal operating data of multiple production sections of the flammable and explosive process production line, use the aforementioned S1 to S5 to generate standard weight data, and generate a standard cloud map based on the standard weight data and the dynamic cloud model; and, based on the current operating data of multiple production sections of the flammable and explosive process production line, use the aforementioned S1 to S5 to generate current weight data, and generate a current dynamic cloud map based on the current weight data and the dynamic cloud model.

[0110] In step S7, the forward cloud transformation and backward cloud transformation algorithms are used to generate the standard cloud map and the dynamic cloud map, as shown in the following example: Figure 3As shown in the figure, a dynamic cloud model is constructed using data from a certain period of time. Standard process operation data is used to form a standard cloud map, and real-time process data is used to form a dynamic cloud map.

[0111] This embodiment uses a cloud model combined with network level analysis results and production line data obtained by sensors to dynamically characterize production line risks. The data sources of the cloud model are divided into two categories: one is static data obtained through network level analysis as a fixed value input, and the other is real-time dynamic data obtained by current and vibration sensors. The two are combined to draw a dynamic cloud map of production line risks, as shown in the effect diagram. Figure 4 shown. Figure 4 In the figure, S1 and S2 are standard cloud images, and S3 and S4 are dynamic cloud images.

[0112] S8, determining the degree of deviation based on the standard cloud map and the current dynamic cloud map, and determining the risk assessment result according to the degree of deviation.

[0113] During specific implementation, the risk level of the process or equipment is judged through intuitive observation or by identifying the degree of deviation between the dynamic cloud map and the standard cloud map through image recognition.

[0114] To leverage cloud model theory and this data to construct a dynamic cloud model, this embodiment uses standard process operation data as a benchmark to form a standard cloud map. Simultaneously, real-time process data dynamically generates a dynamic cloud map. By comparing the degree of deviation between the dynamic cloud map and the standard cloud map, it is possible to intuitively determine whether the current operating status of the process production line or equipment has deviated from normal trajectory, thereby quickly assessing the risk level and providing strong support for timely response measures.

[0115] Traditional cloud model calculation methods only use sensor data or expert scoring data for calculation. The calculation results often deviate from the actual production situation, leading to misjudgment of production line risks, inappropriate risk control measures, and unnecessary economic losses. This embodiment introduces expert experience evaluation values ​​into the cloud model calculation process as an input that is equally important as sensor data. It plays a corrective role in the risk assessment results and achieves the assessment goal of combining subjective and objective risk assessment. It can also be applied to the current situation where most flammable and explosive production lines have a low degree of digitization, lack effective sensor data input, and are highly guided by expert experience.

[0116] In some possible implementations, the established dynamic cloud model is implemented using Python programming, and the compiled program is externally connected to production line equipment and facilities for application. Various sensors installed on the production line can be used to obtain real-time production line dynamic data for real-time evaluation and dynamic analysis.

[0117] In some possible embodiments, after step S8, the following steps (1) to (3) are also included.

[0118] Step (1), determining the contribution data of risk factors based on the risk assessment results;

[0119] Step (2), updating weight data according to contribution data;

[0120] Step (3), update the standard cloud map according to the updated weight data.

[0121] The main innovation of this embodiment lies in the in-depth integration and flexible application of multiple methods to establish a reasonable risk assessment model. The subjective risk judgment matrix of the network level analysis is organically combined with the objective data dynamic calculation of the cloud model to achieve a comprehensive consideration of subjective and objective factors in risk assessment.

[0122] In summary, if Figure 5 As shown, this embodiment first uses the hazard and operability analysis method and the process link review method to extract risk factors and obtain risk factors for multiple risk sections; then, subjective and objective assignments are performed to identify risks in the flammable and explosive product production line from five dimensions: people (subjective, Delphi method), machines (objective, obtained by sensors), materials (subjective, Delphi method), methods (subjective, Delphi method), and environment (objective, obtained by sensors). By organically combining information collected by sensors with expert opinions, a risk quantification assessment method combining subjective and objective factors is implemented; then, the network hierarchical analysis method is used to perform risk weight analysis and obtain weight data; this embodiment is also based on the dynamic representation of risks based on the cloud model. The cloud model data input includes subjective data and objective data, and the cloud model analysis is used to achieve the combination of phased evaluation and real-time evaluation.

[0123] This embodiment starts from the perspective of the risks of flammable and explosive product production line processes, analyzes production line risks using a combination of subjective and objective methods, and proposes a new risk assessment method / model. It uses the hazard and operability analysis method to preliminarily identify production line risks, combines the network hierarchy analysis method to calculate risk factor weights, and uses a cloud model to dynamically characterize production line risks by combining the network hierarchy analysis results and production line data obtained by sensors. Finally, the cloud model is used to compare and analyze standard data with current status data, which makes up for the shortcoming of the original method that it cannot perform real-time dynamic analysis of flammable and explosive production line risks. The dynamic risk assessment results of the cloud model and the network hierarchy analysis results are combined to analyze the comprehensive risk assessment results.

[0124] It should be emphasized that the analytic network process (ANP) has significant advantages in the weighting process and is particularly suitable for complex system decision analysis. Its core advantage is that it can accurately quantify the dynamic interaction and feedback relationship between factors, breaking through the limitations of the traditional analytic hierarchy process (AHP) on the assumption of factor independence. By constructing a networked structural model, the analytic network process not only combines expert experience and quantitative calculations, but also converts qualitative dependencies into multidimensional weights. At the same time, in this embodiment, by using the Delphi method and sensors to obtain real-time data as a pre-input processing method, the subjective influence of "the weight of the dependency relationship between factors depends on expert scoring" caused by the simple use of the analytic network process is eliminated to a certain extent.

[0125] In addition to the aforementioned evaluation method, after step S4, a comprehensive scoring can also be performed, and the comprehensive scoring result is compared with the result obtained in step S8 to determine the final risk assessment result. When performing a comprehensive scoring, first set a scoring standard for each evaluation indicator, such as excellent, good, general, poor, etc., and assign corresponding scores. Then perform score quantification, combine the subjective opinions of experts with objective monitoring values, and use digital means and quantitative evaluation methods and tools such as cloud model theory to quantify the risks. Finally, perform a comprehensive scoring, combine the scores of each indicator with the weights, and obtain a comprehensive safety risk score for the production line. Use the comprehensive safety risk score to compare and analyze the risk assessment results obtained through this embodiment, that is, verify the two results with each other.

[0126] See also Figure 6 An embodiment of the present invention provides a risk assessment device for a flammable and explosive process production line, comprising a data acquisition module 10, a subjective and objective data determination module 20, an assignment module 30, a weighting module 40, a spatiotemporal decomposition module 50, a model building module 60, a quantitative conversion module 70, and a risk assessment module 80.

[0127] The data acquisition module 10 is used to obtain risk factors for multiple production sections of a flammable and explosive process production line. These risk factors are obtained by conducting a risk analysis of these sections using the Hazard and Operability Analysis method. The subjective and objective data determination module 20 collects subjective and objective data for each risk factor from five dimensions: personnel, equipment, raw materials, process methods, and environment, to construct a five-dimensional dataset. The assignment module 30 uses the Delphi method to perform initial risk assignments based on the five-dimensional dataset, obtaining initial risk assignment results. The weighting module 40 uses the analytic network hierarchy process to weight the initial risk assignment results based on the interdependencies and feedback relationships between the risk factors, obtaining weighted data. The spatiotemporal decomposition module 50 performs spatiotemporal decomposition of the production sections of the flammable and explosive process production line with risk points along the time axis, obtaining multiple spatiotemporal segments of data. Each spatiotemporal segment includes weighted data for a specific time period. The model construction module 60 constructs a dynamic cloud model based on the spatiotemporal segment data and cloud model theory. The quantitative conversion module 70 is used to generate standard weight data based on historical normal operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module 10, subjective and objective data determination module 20, assignment module 30, and spatiotemporal decomposition module 50, and to generate a standard cloud map based on the standard weight data and the dynamic cloud model. Furthermore, the quantitative conversion module 70 is used to generate current weight data based on current operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module 10, subjective and objective data determination module 20, assignment module 30, and spatiotemporal decomposition module 50, and to generate a current dynamic cloud map based on the current weight data and the dynamic cloud model. The risk assessment module 80 is used to determine the degree of deviation based on the standard cloud map and the current dynamic cloud map, and to determine the risk assessment result based on the degree of deviation.

[0128] In an alternative embodiment, the subjective data is determined by expert evaluation and the objective data is determined by sensor detection.

[0129] In an optional example, the personnel dimension includes employee quality, operating specifications, and safety responsibilities; the machine and equipment dimension includes equipment status, safety protection, and equipment updates; the raw material dimension includes raw material quality and storage conditions; the process method dimension includes process flow and quality control; and the environment dimension includes production environment, safety facilities, and environmental protection requirements. The personnel dimension, raw material dimension, and process method dimension use expert evaluation to collect subjective data, while the machine and equipment dimension and the environment dimension use sensor equipment to collect objective data.

[0130] In an optional example, the quantitative conversion module 70 generates a standard cloud map and a dynamic cloud map using forward cloud transformation and backward cloud transformation algorithms.

[0131] In an optional example, the assignment module 30 is specifically configured to set a scoring standard for the five-dimensional data set corresponding to each risk factor, and determine an initial assignment based on expert opinions and the scoring standard.

[0132] In an optional example, the model building module 60 includes a membership module, a repetition module, and a parameter generation module.

[0133] The membership module is used to generate a first normal random number Enn, which is a random number about the entropy En of the spatiotemporal segment data. The first normal random number Enn~N(En,He 2 ); where He is the super entropy of the space-time segment data; generates the second normal random number x i , the second normal random number x i is the membership data, the second normal random number x i ~N(Ex,Enn 2 ); where Ex is the expected representation of the spatiotemporal segment data; the membership is determined by the following formula (1):

[0134]

[0135] The repetition module is used to repeatedly execute the above membership module until the total number of cloud droplets is generated.

[0136] The parameter generation module is used to generate cloud model parameters (Ex, En, He) based on the total cloud droplet number and the inverse cloud algorithm (BCT algorithm);

[0137]

[0138] In the above formulas (2), (3) and (4), m is a positive integer, S 2 Represents x i The variance of .

[0139] In an optional embodiment, the risk assessment module 80 includes a contribution data module, a weight update module, and a standard cloud map update module. The contribution data module is used to determine the contribution data of the risk factor based on the risk assessment results, the weight update module is used to update the weight data based on the contribution data, and the standard cloud map update module is used to update the standard cloud map based on the updated weight data.

[0140] Reference Figure 7An embodiment of the present invention further provides an electronic device 1000, including a communication interface 1001, a processor 1002, a memory 1003 and a bus 1004, wherein the processor 1002, the communication interface 1001 and the memory 1003 are connected via the bus 1004; the memory 1003 is used to store a computer program that supports the processor 1002 to execute the above-mentioned flammable and explosive process production line risk assessment method, and the processor 1002 is configured to execute the program stored in the memory 1003.

[0141] Optionally, an embodiment of the present invention further provides a computer-readable medium having a non-volatile program code executable by the processor 1002 , wherein the program code enables the processor 1002 to execute the risk assessment method for flammable and explosive process production lines as described in the above embodiment.

[0142] It is understood from common technical knowledge that the present invention may be implemented by other embodiments that do not depart from its spirit or essential features. Therefore, the embodiments disclosed above are, in all respects, merely illustrative and not exclusive. All modifications within the scope of the present invention or equivalent to the scope of the present invention are intended to be encompassed by the present invention.

Claims

1. A risk assessment method for flammable and explosive process production lines, characterized in that: include: S1, obtaining risk factors for multiple production sections of a flammable and explosive process production line; wherein the risk factors are obtained by performing a risk analysis on the multiple production sections of the flammable and explosive process production line using a hazard and operability analysis method; S2, for each of the risk factors, collect subjective and objective data from five dimensions: personnel dimension, machine and equipment dimension, raw material dimension, process method dimension, and environment dimension to construct a five-dimensional data set; S3, based on the five-dimensional data set, the Delphi method is used to perform initial assignment of risk factors to obtain the initial risk assignment results; S4, based on the mutual dependence and feedback relationship between the risk factors, using the network analysis hierarchy process to weight the initial risk assignment results to obtain weight data; S5, performing spatiotemporal slicing decomposition on the production section with risk points of the flammable and explosive process production line according to the time axis to obtain a plurality of spatiotemporal segment data; wherein each of the spatiotemporal segment data includes the weight data within a specific time period; S6, constructing a dynamic cloud model based on the spatiotemporal segment data and cloud model theory; S7, based on the historical normal operating condition data of multiple production sections of the flammable and explosive process production line, using the aforementioned S1 to S5 to generate standard weight data, and generating a standard cloud map based on the standard weight data and the dynamic cloud model; and based on the current operating condition data of multiple production sections of the flammable and explosive process production line, using the aforementioned S1 to S5 to generate current weight data, and generating a current dynamic cloud map based on the current weight data and the dynamic cloud model; S8, determining a deviation based on the standard cloud map and the current dynamic cloud map, and determining a risk assessment result according to the deviation.

2. The flammable and explosive process production line risk assessment method according to claim 1, characterized in that: The subjective data is determined by expert evaluation, and the objective data is determined by sensor detection.

3. The risk assessment method for flammable and explosive process production lines according to claim 2, characterized in that: The personnel dimension includes employee quality, operating specifications and safety responsibilities; the machine and equipment dimension includes equipment status, safety protection and equipment updates; the raw material dimension includes raw material quality and storage conditions; the process method dimension includes process flow and quality control; and the environment dimension includes production environment, safety facilities and environmental protection requirements; wherein, the personnel dimension, raw material dimension and process method dimension use expert evaluation to collect subjective data, and the machine and equipment dimension and environment dimension use sensor equipment to collect objective data.

4. The flammable and explosive process production line risk assessment method according to claim 1, characterized in that: The S7 includes: The standard cloud image and the dynamic cloud image are generated by using forward cloud transformation and backward cloud transformation algorithms.

5. The flammable and explosive process production line risk assessment method according to claim 2, characterized in that: The S3 includes: A scoring standard is set for the five-dimensional data set corresponding to each risk factor, and the initial assignment is determined in combination with expert opinions and the scoring standard.

6. The flammable and explosive process production line risk assessment method according to claim 1, characterized in that: The S6 includes: Generate a first normal random number Enn, the first normal random number Enn is a random number about the entropy En of the spatiotemporal segment data, the first normal random number Enn~N(En,He 2 ); wherein, He is the super entropy of the space-time segment data; generating a second normal random number x i , the second normal random number x i is the membership degree, the second normal random number x i ~N(Ex,Enn 2 ), i is a positive integer; where Ex is the expectation of the spatiotemporal segment data; the membership degree μ(x i ): Repeat the previous step until the total number of cloud droplets is generated; Based on the total cloud droplet number and the inverse cloud generation algorithm, cloud model parameters (Ex, En, He) are generated; wherein, In the above formulas (2), (3) and (4), m is a positive integer, S 2 Represents x i The variance of .

7. The flammable and explosive process production line risk assessment method according to claim 1, characterized in that: After step S8, the method further includes: Determining contribution data of the risk factors according to the risk assessment results; Updating the weight data according to the contribution data; The standard cloud map is updated according to the updated weight data.

8. A risk assessment device for flammable and explosive process production lines, characterized in that: include: A data acquisition module is used to obtain risk factors of multiple production sections of a flammable and explosive process production line; wherein the risk factors are obtained by performing a risk analysis on the multiple production sections of the flammable and explosive process production line using a hazard and operability analysis method; A subjective and objective data determination module is used to collect subjective and objective data from five dimensions: personnel dimension, machine and equipment dimension, raw material dimension, process method dimension, and environment dimension for each risk factor, and construct a five-dimensional data set; The assignment module is used to perform initial assignment of risk factors based on the five-dimensional data set using the Delphi method to obtain the initial risk assignment results; A weighting module is used to weight the initial risk assignment result based on the mutual dependence and feedback relationship between risk factors using the network analysis method to obtain weight data; A spatiotemporal decomposition module is used to perform spatiotemporal slicing decomposition of production sections with risk points in flammable and explosive process production lines according to the time axis to obtain a plurality of spatiotemporal segment data; wherein each of the spatiotemporal segment data includes the weight data within a specific time period; A model building module, configured to build a dynamic cloud model based on the spatiotemporal segment data and cloud model theory; a quantitative conversion module for generating standard weight data based on historical normal operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module, subjective and objective data determination module, value assignment module, and spatiotemporal decomposition module, and generating a standard cloud map based on the standard weight data and the dynamic cloud model; and, based on current operating condition data of multiple production sections of the flammable and explosive process production line using the aforementioned data acquisition module, subjective and objective data determination module, value assignment module, and spatiotemporal decomposition module, and generating a current dynamic cloud map based on the current weight data and the dynamic cloud model; The risk assessment module is used to determine the deviation degree based on the standard cloud map and the current dynamic cloud map, and determine the risk assessment result according to the deviation degree.

9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images