Sodium hydrosulfite production risk assessment method and system based on multi-source data fusion

Abstract

The invention relates to a sodium hydrosulfite production risk assessment method and system based on multi-source data fusion. The system comprises a production process data acquisition module, a key variable linkage analysis module, a real-time risk detection module, a dynamic risk trigger analysis module and a comprehensive risk score calculation module. The method not only improves the risk assessment precision in the sodium hydrosulfite production process, but also realizes effective risk early warning and control through real-time and dynamic monitoring and regulation means. Meanwhile, the method fully combines the particularity of the sodium hydrosulfite production process, has significant value in safety guarantee, production efficiency improvement and management optimization, and establishes a new risk assessment normal form with scientificity and practicability for the chemical industry.

Description

Technical Field

[0001] The present invention relates to the technical field of production evaluation, and specifically to a method and system for evaluating the production risk of sodium dithionite based on multi-source data fusion. Background Art

[0002] The technical field of sodium dithionite production evaluation mainly focuses on the comprehensive analysis and optimization of its production process, aiming to improve production efficiency, product quality, and environmental protection performance. This field covers aspects such as raw material selection, control and optimization of chemical reaction processes, improvement of product purity, energy consumption management, and treatment of waste gas and wastewater. Through scientific technical means, the economy and sustainability of the production process are evaluated, while reducing the risks of resource waste and environmental pollution. In addition, this field also pays attention to the stable operation of equipment, the introduction of new technologies and their promoting effects on production efficiency, ensuring the safety and high efficiency of the entire production process. Through comprehensive evaluation, technical guidance and optimization solutions are provided for enterprises, promoting the industry to develop in the direction of environmental protection, high efficiency, and sustainability.

[0003] There are many problems to be solved in the existing technology for evaluating the production risk of sodium dithionite. First of all, the complex linkage relationships between variables have not been fully explored, and traditional methods have insufficient analysis capabilities for the causal relationships and dynamic linkage effects of multiple variables, resulting in the inability to accurately identify the key factors forming risks. In addition, the distinction between chronic cumulative risks and sudden risks and the tracking of dynamic evolution trends are also insufficient, making it difficult to formulate precise countermeasures for different types of risks. At the same time, the adaptability of the fixed threshold model is poor, and it is difficult to dynamically adjust under complex production conditions, prone to false alarms or missed alarms. The existing technology has not fully integrated multi-source data, resulting in limited comprehensiveness and accuracy of the risk assessment model, and lacking the ability of quantitative analysis and intuitive expression of the risk status. In addition, the visualization annotation ability of the spatial distribution of risk variables and high-risk areas is also relatively weak, making it difficult to provide intuitive and clear decision-making support for managers. Summary of the Invention

[0004] In view of the technical problems existing in the prior art, the present invention provides a method and system for evaluating the production risk of sodium dithionite based on multi-source data fusion.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: A method for evaluating the production risk of sodium dithionite based on multi-source data fusion, the method comprising:

[0006] Collect production process data, including environmental data, production process data, equipment data, and operation behavior data, add time series tags to the data, and preprocess the collected data;

[0007] Based on the characteristics of sodium dithionite production, a multivariate causal association model was constructed to analyze the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and to capture the linkage effect of key variables;

[0008] Combined with production process data, a detection model is established to monitor and classify the current risk status in real time, including normal status, chronic cumulative risk and sudden risk. At the same time, the trend of risk change is identified to obtain the risk accumulation index and sudden risk index.

[0009] Set dynamic risk trigger thresholds based on causal relationships, analyze the correlation between the risk accumulation index and the sudden risk index and the dynamic risk trigger thresholds, and dynamically update the monitoring strategy;

[0010] Analyze the relationship between the risk accumulation index and the sudden risk index, and calculate the comprehensive risk score of the production line based on the linkage relationship strength, risk accumulation index and sudden risk index, and visually mark the risk situation and risk degree.

[0011] As a further solution of the present invention, the analysis of the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and capturing the linkage effect of key variables, specifically includes:

[0012] Analyze the characteristics of sodium bisulfite and identify the cause-effect relationship between sodium bisulfite production and humidity, dust concentration, equipment status and reaction temperature;

[0013] Based on whether the parameters used in the causal relationship analysis are key variables, an initial causal model is constructed in combination with historical production data;

[0014] Combined with the initial causal model, path analysis is used to calculate the strength of the linkage relationship between variables:

[0015] Based on the strength of the linkage relationship between variables, identify the key linkage effects that have the greatest impact on the risk status.

[0016] As a further solution of the present invention, the strength of the linkage relationship between the calculated variables is specifically:

[0017] For directly linked variables:

[0018] Among them, represents the linkage strength between variable i and variable j, which is the path coefficient, including: the path coefficient P of humidity on dust concentration H , Path coefficient P of dust concentration on explosion risk D , the path coefficient P of equipment status on leakage risk E and the path coefficient P of reaction temperature to dust concentration T , σ i and σ jThe standard deviations of variable i and variable j respectively;

[0019] For variables with indirect paths:

[0020] Among them, represents the total linkage relationship strength between variable i and variable j, P ik represents the path coefficient of variable i to intermediate variable k, and represents the linkage relationship strength of intermediate variable k to variable j;

[0021] The identification of the key linkage effect with the greatest impact on the risk state is specifically:

[0022] C ij = S ij ·W i ;

[0023] Among them, C ij is the risk contribution rate of variable i and variable j, W i is the sensitivity weight of variable j.

[0024] As a further solution of the present invention, the identification of the risk change trend, obtaining the risk accumulation index and the sudden risk index specifically includes:

[0025] Based on real-time data, the production state is divided into different risk levels, including:

[0026] Normal state, all variables are within the safe range and there is no abnormal signal;

[0027] Cumulative risk, the variable is close to the limit value and the variable has a progressive nature;

[0028] Sudden risk, the variable grows abnormally and is close to the limit value, exceeding the critical value;

[0029] Establish a detection model to quantify the risk change trend and calculate the risk accumulation index and the sudden risk index:

[0030]

[0031] Among them, RAI is the risk accumulation index, V i is the risk weight of variable i, f(X i ) is the deviation degree of variable i;

[0032]

[0033] Among them, FRI is the sudden risk index, is the instantaneous change rate of variable i;

[0034] Establish a time series model to predict the change trend of the future risk accumulation index.

[0035] As a further solution of the present invention, the establishment of the time series model to predict the change trend of the future risk accumulation index is specifically as follows:

[0036] RAI(t + Δt) = RAI(t) + ΔRAI(t);

[0037]

[0038] wherein, RAI(t + Δt) is the expected risk accumulation index at the future moment t + Δt, RAI(t) is the risk accumulation index at the current moment t, and ΔRAI(t) is the change amount of the risk accumulation index within the time interval from t to t + Δt, X i (t) is the actual value of the variable i at the current moment, is the change rate of the variable i, and γ i is the accumulation effect coefficient.

[0039] As a further solution of the present invention, the setting of the dynamic risk trigger threshold, and the analysis of the association between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and the dynamic update of the monitoring strategy specifically include:

[0040] Based on the production process and historical data, construct an initial risk trigger threshold range for each variable, and adjust to obtain the final risk trigger threshold range based on whether the variable is a key variable;

[0041] Calculate the triggering relationship between the risk accumulation index and the risk trigger threshold:

[0042] If the risk index has exceeded the trigger threshold, adjust all variables in combination with the strength of the linkage relationship;

[0043] Calculate the triggering relationship between the sudden risk index and the risk trigger threshold:

[0044] If the risk change rate has exceeded the trigger threshold, adjust all variables in combination with the strength of the linkage relationship.

[0045] As a further solution of the present invention, the calculation of the triggering relationship between the risk accumulation index and the risk trigger threshold is specifically as follows:

[0046]

[0047] wherein, D RAI is the deviation degree of the risk index, and T RAI is the trigger threshold of the risk accumulation index;

[0048] If D RAI > 0, it means that the risk index has exceeded the trigger threshold, and all variables are adjusted in combination with the strength of the linkage relationship;

[0049] The triggering relationship between the calculated sudden risk index and the risk trigger threshold is specifically as follows:

[0050]

[0051] Among them, R FRI is the change rate of the sudden risk index;

[0052] If it means that the risk change rate has exceeded the trigger threshold, and all variables are adjusted in combination with the linkage relationship strength.

[0053] As a further solution of the present invention, analyzing the relationship between the risk accumulation index and the sudden risk index, and calculating the comprehensive risk score of the production line based on the linkage relationship strength, the risk accumulation index and the sudden risk index, specifically including:

[0054] Calculating the time series correlation between the risk accumulation index and the sudden risk index:

[0055] ρ RAI,FRI (t) = Corr(RAI(t), FRI(t + Δt'));

[0056] Among them, ρ RAI,FRI (t) is the correlation coefficient between the risk accumulation index and the sudden risk index at the current time t, and Δt' represents the lag time of the sudden risk relative to the accumulated risk;

[0057] Based on multi-source data and multi-dimensional risk indicators, constructing a risk scoring model for comprehensively evaluating the safety status of the production line, and calculating the comprehensive risk assessment result of the overall production;

[0058] Identifying all relevant variables that cause risks, and establishing a spatial distribution map, and marking the spatial distribution map according to the variable detection positions.

[0059] As a further solution of the present invention, the calculation of the comprehensive risk assessment result of the overall production is specifically as follows:

[0060] R 总 = W RAI ·RAI + W FRI ·FRI + W ρ ·ρ RAI,FRI (t) + W S ·∑ i,j S ij ;

[0061] Among them, R 总 is the comprehensive risk score, W ρ , W S , W RAI , WFRI The correlation coefficients, linkage relationship strengths, and weight coefficients of the risk accumulation index and the sudden risk index, respectively.

[0062] Another object of the present invention is to provide a sodium hydrosulfite production risk assessment system based on multi-source data fusion. The system includes:

[0063] A production process data acquisition module for acquiring production process data, including environmental data, production process data, equipment data, and operation behavior data, adding time series tags to the data, and preprocessing the acquired data;

[0064] A key variable linkage analysis module for constructing a multi-variable causal association model based on the characteristics of sodium hydrosulfite production, analyzing the causal relationship between humidity, dust concentration, equipment status, and reaction temperature, and capturing the linkage effect of key variables;

[0065] A real-time risk detection module for establishing a detection model in combination with production process data, real-time monitoring and grading the current risk status, including normal status, chronic cumulative risk, and sudden risk, identifying the risk change trend at the same time, and obtaining the risk accumulation index and the sudden risk index;

[0066] A dynamic risk trigger analysis module for setting a dynamic risk trigger threshold in combination with the causal relationship, analyzing the association between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and dynamically updating the monitoring strategy;

[0067] A comprehensive risk score calculation module for analyzing the relationship between the risk accumulation index and the sudden risk index, calculating the comprehensive risk score of the production line based on the linkage relationship strength, the risk accumulation index, and the sudden risk index, and visually annotating the risk situation and risk level.

[0068] The beneficial effects of the present invention are:

[0069] By combining a multi-variable causal association model and a real-time detection mechanism, this method comprehensively captures the complex linkage effects among key variables such as humidity, dust concentration, equipment status, and reaction temperature, and deeply explores the causal relationship between them and risk formation. This method effectively solves problems such as diverse risk sources and high coupling strength among variables in the production environment, making risk assessment more accurate and reducing the possibility of misjudgment and omission. At the same time, using the risk accumulation index and the sudden risk index, the system can dynamically monitor the evolution trends of chronic cumulative risks (such as equipment aging) and sudden risks (such as drastic changes in humidity), achieving comprehensive hierarchical control of risks of different natures and enhancing the comprehensiveness and pertinence of the assessment.

[0070] The setting of the dynamic risk trigger threshold further enhances the flexibility and responsiveness of the system. By combining the linkage strength of causal analysis and the real-time risk status, the system can dynamically adjust the monitoring strategy according to the real-time production situation, avoiding the problems of "over-warning" or lagged response that may be caused by traditional fixed thresholds. This dynamic adjustment mechanism is particularly suitable for the highly sensitive characteristics of sodium hydrosulfite production to environmental variables, ensuring that precise intervention can be achieved at an early stage of risk control and reducing the likelihood of accidents. In addition, the comprehensive risk scoring model presents the risk analysis results in a quantitative form by integrating multi-dimensional data, and intuitively identifies risk hotspots by associating the temporal dynamics with the spatial distribution map, further optimizing the operability and decision-making efficiency of risk management. Brief Description of the Drawings

[0071] Figure 1 It is a flowchart of the method for risk assessment of sodium hydrosulfite production based on multi-source data fusion provided by an embodiment of the present invention;

[0072] Figure 2 It is a flowchart of analyzing the causal relationship among humidity, dust concentration, equipment status and reaction temperature, and capturing the linkage effect of key variables provided by an embodiment of the present invention;

[0073] Figure 3 It is a flowchart of identifying the risk change trend, obtaining the risk accumulation index and the sudden risk index provided by an embodiment of the present invention;

[0074] Figure 4 It is a flowchart of setting the dynamic risk trigger threshold, analyzing the relationship between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and dynamically updating the monitoring strategy provided by an embodiment of the present invention;

[0075] Figure 5 It is a flowchart of analyzing the relationship between the risk accumulation index and the sudden risk index, and calculating the comprehensive risk score of the production line based on the linkage relationship strength, the risk accumulation index and the sudden risk index provided by an embodiment of the present invention;

[0076] Figure 6 It is a structural block diagram of the risk assessment system for sodium hydrosulfite production based on multi-source data fusion provided by an embodiment of the present invention. Detailed Embodiments

[0077] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0078] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0079] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0080] Figure 1 A flow chart of a sodium dithionite production risk assessment method based on multi-source data fusion provided in an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method includes:

[0081] S100, collecting production process data, including environmental data, production process data, equipment data, and operation behavior data, adding time series labels to the data, and preprocessing the collected data;

[0082] In terms of data collection and processing, in order to accurately capture key information that may cause risks in the production process of sodium dithionite, it is necessary to comprehensively cover multiple dimensions of data in the production process, including environmental data (such as humidity, dust concentration), production process data (such as reaction temperature, reaction time), equipment data (such as equipment operating status, pipeline pressure), and operation behavior data (such as personnel operation records, abnormal behavior logs). At the same time, in order to ensure the timeliness and accuracy of the data, time series labels need to be added to the data for subsequent analysis. In addition, the collected raw data also needs to be systematically preprocessed, including denoising, outlier removal, data smoothing, etc., to ensure the quality and reliability of the data.

[0083] S200, based on the characteristics of sodium dithionite production, builds a multivariate causal association model to analyze the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and captures the linkage effect of key variables;

[0084] Since sodium hydrosulfite highly depends on the precise control of chemical reactions during the production process, its sensitivity to environmental factors is significantly higher than that of other chemical products. For example, humidity not only directly affects the stability of the reaction but may also cause changes in the hygroscopicity of dust particles, thus increasing the risk of dust explosion in the system; changes in dust concentration may indirectly affect the temperature fluctuations of the reaction through equipment wear or blockage. In addition, the equipment status (such as wear or aging) may exacerbate the unevenness of reactant input, thereby further increasing the perturbation of temperature fluctuations on chemical reactions. Therefore, through causal relationship analysis, the linkage mechanism among these variables can be accurately identified.

[0085] Secondly, based on the results of causal relationship analysis and combined with historical production data, an initial causal model is constructed. This model adopts multi-source heterogeneous data fusion technology to uniformly model data from different sources (such as sensor data, equipment operation logs, environmental monitoring data) to ensure the scientificity and comprehensiveness of the causal model. On this basis, using path analysis methods, the strength of the linkage relationship between humidity, dust concentration, equipment status and reaction temperature is quantified. In the path analysis calculation formula, the weights of each path coefficient (such as the path coefficient of humidity on dust concentration, the path coefficient of reaction temperature on explosion probability) are clarified, and the influence of the dimensional differences of different variables is eliminated through standardization processing, making the causal association model more general and applicable.

[0086] Finally, by comprehensively analyzing the direct and indirect relationships among variables, the key linkage effects that have the greatest impact on the risk status are further identified. This process not only focuses on the risk contribution of a single variable but also considers the complex interaction among variables. For example, the amplification effect of the triple linkage of humidity, dust concentration and equipment status on sudden risks. By calculating the total strength of the variable linkage relationship and combining the sensitivity weights of each variable, the contribution degree of key variables to the overall risk can be quantified, and finally the parameters that need to be monitored and controlled with the highest priority can be determined.

[0087] Through causal relationship modeling, this step can deeply explore the complex dynamic associations among multiple variables, providing a scientific basis for risk early warning. This method is not limited to the static monitoring of a single variable but combines the linkage effects among variables, providing a more comprehensive perspective for risk prediction. Secondly, the introduction of the path analysis method enables the causal model to not only explain the strength of the association between variables but also quantify the degree of impact of specific paths on risks, thus providing precise guidance for dynamically adjusting the production process and optimizing the monitoring strategy. In addition, by identifying the key linkage effects, it is possible to effectively concentrate limited resources on focusing on monitoring and controlling the factors that have the greatest impact on production risks, thereby reducing the monitoring cost and improving the efficiency of overall risk management. Finally, this process fully considers the particularities of sodium hydrosulfite production, such as the sensitivity characteristics of environmental humidity and dust concentration, making the risk assessment results more targeted and practical, providing important technical support for enhancing the safety and stability of chemical production.

[0088] As Figure 2 shown, analyzing the causal relationships among humidity, dust concentration, equipment status, and reaction temperature and capturing the linkage effects of key variables specifically include:

[0089] S210, analyze the characteristics of sodium hydrosulfite and identify the causal relationships between sodium hydrosulfite production and humidity, dust concentration, equipment status, and reaction temperature;

[0090] S220, based on whether the parameters used in the causal relationship analysis are key variables, combined with historical production data, construct an initial causal model;

[0091] S230, combined with the initial causal model, use path analysis to calculate the strength of the linkage relationship between variables:

[0092] S240, based on the strength of the linkage relationship between variables, identify the key linkage effects that have the greatest impact on the risk status.

[0093] In this step, the calculation of the strength of the linkage relationship between variables is specifically:

[0094] For directly related variables:

[0095] where represents the strength of the linkage relationship between variable i and variable j, and is the path coefficient, including: the path coefficient P of humidity on dust concentration H , the path coefficient P of dust concentration on explosion risk D , the path coefficient P of equipment status on leakage risk E and the path coefficient P of reaction temperature on dust concentration T , σ i and σ jThe standard deviations of variable i and variable j respectively;

[0096] For variables with indirect paths:

[0097] Among them, represents the total linkage relationship strength between variable i and variable j, P ik represents the path coefficient of variable i to intermediate variable k, and represents the linkage relationship strength of intermediate variable k to variable j;

[0098] The identification of the key linkage effect with the greatest impact on the risk state is specifically:

[0099] C ij = S ij ·W i ;

[0100] Among them, C ij is the risk contribution rate of variable i and variable j, and W i is the sensitivity weight of variable j.

[0101] S300, combined with production process data, establish a detection model to monitor and classify the current risk state in real time, including normal state, chronic cumulative risk and sudden risk, and at the same time identify the risk change trend, and obtain the risk accumulation index and sudden risk index;

[0102] In this step, the production state is divided into three risk levels: normal state, cumulative risk and sudden risk. Among them, the normal state indicates that all variables are within the safe range, and the production is in a stable operation stage at this time; the cumulative risk reflects that the variables are gradually approaching the limit value and show a progressive change, such as equipment fatigue or chronic deterioration of environmental parameters; the sudden risk focuses on the rapid and abnormal growth of variables, such as the drastic fluctuation of humidity or dust concentration, usually accompanied by the situation of exceeding the safety threshold. Such a grading method realizes the precise characterization of the risk state in different dimensions and provides a clear judgment basis for subsequent risk control.

[0103] By quantifying the risk change trend, calculate the risk accumulation index and sudden risk index. The risk accumulation index characterizes the long-term cumulative risk by integrating the risk weights of each variable in the production process and the deviation degree function of the variable, and reflects the trend of the variable gradually deviating from the safe range. The sudden risk index quantifies the suddenness and unexpectedness of the risk by capturing the change characteristics of the instantaneous change rate of the variable. The combination of the two indexes can comprehensively describe the superposition effect of chronic risks and unexpected events in the production process and further improve the sensitivity of risk monitoring.

[0104] By performing time - series modeling on historical data to predict the changing trend of the future risk accumulation index, it can not only provide a quantitative early warning of future risks, but also provide a scientific basis for production managers to optimize production parameters and take intervention measures in advance.

[0105] As Figure 3 shown, the identification of the risk change trend, obtaining the risk accumulation index and the sudden - risk index specifically includes:

[0106] S310, dividing the production status into different risk levels based on real - time data, including:

[0107] Normal state, all variables are within the safe range and there are no abnormal signals;

[0108] Cumulative risk, the variable is close to the limit value and the variable has a progressive nature;

[0109] Sudden risk, the variable has an abnormal increase and is close to the limit value and exceeds the critical value;

[0110] S320, establishing a detection model to quantify the risk change trend and calculate the risk accumulation index and the sudden - risk index:

[0111]

[0112] Among them, RAI is the risk accumulation index, V i is the risk weight of variable i, f(X i ) is the deviation degree of variable i;

[0113]

[0114] Among them, FRI is the sudden - risk index, is the instantaneous change rate of variable i;

[0115] S330, establishing a time - series model to predict the changing trend of the future risk accumulation index.

[0116] In this step, the establishment of the time - series model to predict the changing trend of the future risk accumulation index is specifically:

[0117] RAI(t + Δt)=RAI(t)+ΔRAI(t);

[0118]

[0119] Among them, RAI(t + Δt) is the expected risk accumulation index at the future moment t + Δt, RAI(t) is the risk accumulation index at the current moment t, ΔRAI(t) is the change amount of the risk accumulation index within the time interval from t to t + Δt, X i(t) is the actual value of variable i at the current moment, is the change rate of variable i, γ i is the cumulative effect coefficient.

[0120] S400, set a dynamic risk trigger threshold in combination with the causal relationship, and analyze the association between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and dynamically update the monitoring strategy;

[0121] This step constructs an initial risk trigger threshold range for different variables (such as humidity, dust concentration, equipment status, reaction temperature, etc.). This process needs to comprehensively consider the safety requirements of the production process and the historical data law to ensure that the initial threshold can reflect the normal production fluctuation range. Since the influence degrees of various variables in the sodium hydrosulfite production are different, the key variables can be identified through causal association analysis, and then the trigger threshold range thereof is adjusted. The threshold adjustment of the key variables needs to be more strict in order to capture potential risk signals at an early stage, while the threshold of non-key variables can be appropriately relaxed to avoid over-warning and improve the robustness and practicability of the system.

[0122] Secondly, based on the trigger relationship of the risk accumulation index, dynamically evaluate whether the current risk status has exceeded the trigger threshold. When the risk accumulation index is greater than 0, it indicates that the chronic risks accumulated in the long term (such as equipment aging or small offsets in process conditions) have reached the level that requires intervention. In this case, by combining the strength of the linkage relationship (such as the interaction effect between humidity and dust concentration), the threshold range of all variables is adjusted in real time. This adjustment process can accurately capture the systematic changes brought about by risk accumulation, so as to implement effective intervention before the chronic risk gradually evolves into an obvious risk.

[0123] At the same time, by calculating the change rate of the sudden risk index and its trigger relationship, it is possible to quickly respond to sudden events that may occur in the production process. For example, when the humidity suddenly rises sharply or a certain component of the equipment fails, the change rate of the sudden risk index exceeds the set threshold, and the system will immediately trigger the warning mechanism and dynamically adjust the monitoring strategies of all variables in combination with the strength of the linkage relationship between variables. This method enables the production management system not only to monitor the cumulative effect of risks, but also to quickly respond when sudden risks occur, forming a double guarantee.

[0124] By combining the characteristics of chronic risk accumulation and sudden risk changes, the system can adapt to risk dynamics at different time scales, avoiding misjudgment or lagged response problems that may be caused by fixed thresholds. In addition, the practice of adjusting the variable monitoring strategy based on the intensity of the linkage relationship helps to comprehensively capture the complex interactions of various variables in sodium dithionite production. Especially for systematic risks caused by multi-variable coupling, the dynamic adjustment mechanism can provide more targeted risk control measures, effectively reducing the probability of accidents and optimizing production efficiency.

[0125] As Figure 4 shown, setting the dynamic risk trigger threshold and analyzing the association between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and dynamically updating the monitoring strategy specifically includes:

[0126] S410, based on the production process and historical data, construct an initial risk trigger threshold range for each variable, and adjust to obtain the final risk trigger threshold range based on whether the variable is a key variable;

[0127] S420, calculate the triggering relationship of the risk accumulation index to the risk trigger threshold:

[0128] S430, if the risk index has exceeded the trigger threshold, and adjust all variables in combination with the intensity of the linkage relationship;

[0129] S440, calculate the triggering relationship of the sudden risk index to the risk trigger threshold:

[0130] S450, if the risk change rate has exceeded the trigger threshold, adjust all variables in combination with the intensity of the linkage relationship.

[0131] In this step, the calculation of the triggering relationship of the risk accumulation index to the risk trigger threshold is specifically:

[0132]

[0133] Among them, D RAI is the deviation degree of the risk index, and T RAI is the trigger threshold of the risk accumulation index;

[0134] If D RAI > 0, it means that the risk index has exceeded the trigger threshold, and adjust all variables in combination with the intensity of the linkage relationship;

[0135] The calculation of the triggering relationship of the sudden risk index to the risk trigger threshold is specifically:

[0136]

[0137] Among them, R FRI is the change rate of the sudden risk index;

[0138] If it indicates that the risk change rate has exceeded the trigger threshold, and all variables are adjusted in combination with the intensity of the linkage relationship.

[0139] In S500, analyze the relationship between the risk accumulation index and the sudden risk index, and calculate the comprehensive risk score of the production line based on the intensity of the linkage relationship, the risk accumulation index, and the sudden risk index, and visually label the risk situation and risk level.

[0140] This step calculates the time series correlation between the risk accumulation index and the sudden risk index. This step not only reveals the time series linkage relationship between chronic risk and sudden risk, but also determines the lagging impact of sudden risk on chronic risk accumulation by analyzing time delay. This kind of analysis can effectively capture the dynamic characteristics of risk evolution and provide a scientific basis for the subsequent comprehensive risk scoring model.

[0141] In the comprehensive risk scoring model, by introducing multi-dimensional weight parameters, key indicators such as the risk accumulation index, the sudden risk index, the time correlation, and the intensity of the variable linkage relationship are comprehensively weighted and calculated to obtain the comprehensive risk score of the overall production line. This scoring process fully considers the particularity of sodium hydrosulfite production, such as its high sensitivity to humidity, dust concentration, and reaction temperature, as well as the complex interaction between these variables. By assigning different weights to each indicator, the system can dynamically adjust the influence degree of key variables, making the scoring result more in line with the actual production scenario.

[0142] In addition, S500 also particularly emphasizes the construction and annotation of the spatial distribution map. By identifying all relevant variables that cause risks and combining their monitoring location information, an intuitive spatial distribution map is established. This can not only clearly present high-risk areas and key risk points, but also provide production managers with an intuitive risk distribution situation, which helps to carry out targeted risk intervention. For example, if the state of a certain device has a significant linkage impact on the surrounding environment (such as humidity and dust concentration), the problem can be quickly located through the spatial distribution map, and the operation or maintenance strategy of the relevant area can be optimized.

[0143] The significant advantage of this step is that its integrity and comprehensiveness significantly improve the accuracy and practicality of risk assessment. First, the introduction of time series correlation quantifies the dynamic linkage relationship between chronic risk and sudden risk, which provides an important basis for risk prediction and early warning. At the same time, the comprehensive risk scoring model unifies and quantifies different types of risk indicators through multi-dimensional weighting, ensuring the objectivity and comprehensiveness of the evaluation results. Second, the visual annotation of the spatial distribution map makes the risk analysis results more intuitive, provides clear operation guidelines for the risk management of the production line, and helps to quickly and accurately take targeted measures.

[0144] This step combines the particularities of sodium dithionite production, such as its sensitive production environment requirements and complex variable interaction effects, ensuring a high degree of applicability of the risk assessment model to the actual process. This not only optimizes the operating efficiency of the production line but also greatly reduces the probability of safety accidents, providing a reference example for risk assessment and management in the entire chemical industry.

[0145] As Figure 5 shown, the relationship between the analyzed risk accumulation index and the sudden risk index is examined, and based on the linkage relationship strength, risk accumulation index, and sudden risk index, the comprehensive risk score of the production line is calculated, specifically including:

[0146] S510, calculate the time series correlation between the risk accumulation index and the sudden risk index:

[0147] ρ RAI,FRI (t) = Corr(RAI(t), FRI(t + Δt′));

[0148] where ρ RAI,FRI (t) is the correlation coefficient between the risk accumulation index and the sudden risk index at the current time t, and Δt′ represents the lag time of the sudden risk relative to the accumulated risk;

[0149] S520, construct a risk score model for comprehensively evaluating the safety status of the production line based on multi-source data and multi-dimensional risk indicators, and calculate the comprehensive risk assessment result of the overall production;

[0150] S530, identify all relevant variables that cause risks, establish a spatial distribution map, and mark the spatial distribution map according to the variable detection positions.

[0151] In this step, the calculation of the comprehensive risk assessment result of the overall production is specifically as follows:

[0152] R 总 = W RAI ·RAI + W FRI ·FRI + W ρ ·ρ RAI,FRI (t) + W S ·∑ i,j S ij ;

[0153] where R 总 is the comprehensive risk score, and W ρ , W S , W RAI , W FRI are the weight coefficients of the correlation coefficient between the risk accumulation index and the sudden risk index, the linkage relationship strength, the risk accumulation index, and the sudden risk index, respectively.

[0154] Figure 6 This is the structural block diagram of the sodium dithionite production risk assessment system based on multi-source data fusion provided by the embodiments of the present invention. As Figure 6 shown, the system includes:

[0155] A production process data acquisition module 100, which is used to collect production process data, including environmental data, production process data, equipment data, and operation behavior data, add time series tags to the data, and preprocess the collected data;

[0156] A key variable linkage analysis module 200, which is used to build a multi-variable causal association model based on the characteristics of sodium dithionite production, analyze the causal relationship between humidity, dust concentration, equipment status, and reaction temperature, and capture the linkage effect of key variables;

[0157] A real-time risk detection module 300, which is used to establish a detection model in combination with production process data, monitor and classify the current risk status in real time, including normal status, chronic cumulative risk, and sudden risk, identify the risk change trend at the same time, and obtain the risk accumulation index and sudden risk index;

[0158] A dynamic risk trigger analysis module 400, which is used to set a dynamic risk trigger threshold in combination with the causal relationship, analyze the association between the risk accumulation index and the sudden risk index and the dynamic risk trigger threshold, and dynamically update the monitoring strategy;

[0159] A comprehensive risk score calculation module 500, which is used to analyze the relationship between the risk accumulation index and the sudden risk index, calculate the comprehensive risk score of the production line based on the linkage strength, risk accumulation index, and sudden risk index, and visually label the risk situation and risk level.

[0160] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0161] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0162] The present invention will be described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce a means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0163] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0164] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0165] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concepts. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0166] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A sodium dithionite production risk assessment method based on multi-source data fusion, characterized in that: The method comprises: Collect production process data, including environmental data, production process data, equipment data, and operation behavior data, add time series labels to the data, and pre-process the collected data; Based on the characteristics of sodium dithionite production, a multivariate causal association model was constructed to analyze the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and to capture the linkage effect of key variables; Combined with production process data, a detection model is established to monitor and classify the current risk status in real time, including normal status, chronic cumulative risk and sudden risk. At the same time, the trend of risk change is identified to obtain the risk accumulation index and sudden risk index. Set dynamic risk trigger thresholds based on causal relationships, analyze the correlation between the risk accumulation index and the sudden risk index and the dynamic risk trigger thresholds, and dynamically update the monitoring strategy; Analyze the relationship between the risk accumulation index and the sudden risk index, and calculate the comprehensive risk score of the production line based on the linkage relationship strength, risk accumulation index and sudden risk index, and visually mark the risk situation and risk degree.

2. The method according to claim 1, characterized in that The analysis of the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and capturing the linkage effects of key variables, specifically includes: Analyze the characteristics of sodium bisulfite and identify the cause-effect relationship between sodium bisulfite production and humidity, dust concentration, equipment status and reaction temperature; Based on whether the parameters used in the causal relationship analysis are key variables, an initial causal model is constructed in combination with historical production data; Combined with the initial causal model, path analysis is used to calculate the strength of the linkage relationship between variables: Based on the strength of the linkage relationship between variables, identify the key linkage effects that have the greatest impact on the risk status.

3. The method according to claim 2, characterized in that The strength of the linkage relationship between the calculated variables is specifically: For directly linked variables: Among them, represents the linkage strength between variable i and variable j, is the path coefficient, including: the path coefficient P of humidity on dust concentration H , Path coefficient P of dust concentration on explosion risk D , the path coefficient P of equipment status on leakage risk E and the path coefficient P of reaction temperature to dust concentration T , σ i and σ j are the standard deviations of variables i and j respectively; For variables with indirect paths: in, represents the total linkage strength between variable i and variable j, P ik It represents the path coefficient of variable i to intermediate variable k, and it represents the linkage strength of intermediate variable k to variable j; The identification of the key linkage effects that have the greatest impact on the risk status is specifically: C ij =S ij ·W i ; Among them, C ij is the risk contribution ratio of variable i and variable j, W i is the sensitivity weight of variable j.

4. The method according to claim 2, characterized in that: The identifying of the risk change trend and obtaining the risk accumulation index and the sudden risk index specifically include: The production status is divided into different risk levels based on real-time data, including: Normal state, all variables are within the safe range, and there are no abnormal signals; Cumulative risk, the variable is close to the limit value, and the variable is gradual; Sudden risk: abnormal growth of variables approaching limit values ​​or exceeding critical values; Establish a detection model, quantify the risk change trend, and calculate the risk accumulation index and sudden risk index: Among them, RAI is the risk accumulation index, V i is the risk weight of variable i, f(X i ) is the degree of deviation of variable i; Among them, FRI is the sudden risk index, is the instantaneous rate of change of variable i; Establish a time series model to predict the changing trend of the future risk accumulation index.

5. The method according to claim 4, characterized in that The time series model is established to predict the changing trend of the future risk accumulation index, specifically: RAI(t+Δt)=RAI(t)+ΔRAI(t); Among them, RAI(t+Δt) is the expected risk accumulation index at the future time t+Δt, RAI(t) is the risk accumulation index at the current time t, ΔRAI(t) is the change in the risk accumulation index in the time interval from t to t+Δt, and X i (t) is the actual value of variable i at the current moment, is the rate of change of variable i, γ i is the accumulation effect coefficient.

6. The method according to claim 4, characterized in that The setting of dynamic risk triggering thresholds, analyzing the correlation between the risk accumulation index and the sudden risk index and the dynamic risk triggering thresholds, and dynamically updating the monitoring strategy specifically include: Based on the production process and historical data, an initial risk trigger threshold range is constructed for each variable, and the final risk trigger threshold range is adjusted based on whether the variable is a key variable; Calculate the trigger relationship between the risk accumulation index and the risk trigger threshold: If the risk index exceeds the trigger threshold, all variables will be adjusted based on the strength of the linkage relationship; Calculate the trigger relationship between the sudden risk index and the risk trigger threshold: If the risk change rate exceeds the trigger threshold, all variables will be adjusted based on the strength of the linkage relationship.

7. The method according to claim 6, characterized in that The trigger relationship between the calculated risk accumulation index and the risk trigger threshold is specifically: Among them, D RAI is the risk index deviation degree, T RAI is the trigger threshold of the risk accumulation index; If D RAI >0, it means that the risk index has exceeded the trigger threshold, and all variables are adjusted according to the linkage strength; The trigger relationship between the sudden risk index and the risk trigger threshold is calculated as follows: Among them, R FRI is the rate of change of the sudden risk index; like This means that the risk change rate has exceeded the trigger threshold, and all variables are adjusted based on the strength of the linkage relationship.

8. The method according to claim 6, characterized in that The analysis of the relationship between the risk accumulation index and the sudden risk index, and calculation of the comprehensive risk score of the production line based on the linkage relationship strength, the risk accumulation index and the sudden risk index, specifically includes: Calculate the time series correlation between the risk accumulation index and the sudden risk index: ρ RAI,FRI (t)=Corr(RAI(t),FRI(t+Δt) ′ )); Among them, ρ RAI,FRI (t) is the correlation coefficient between the risk accumulation index and the sudden risk index at the current moment t, Δt ′ It indicates the lag time of sudden risk relative to accumulated risk; Based on multi-source data and multi-dimensional risk indicators, a risk scoring model is constructed to comprehensively evaluate the safety status of the production line and calculate the comprehensive risk assessment results of the overall production; Identify all relevant variables that cause risks, establish a spatial distribution map, and mark the spatial distribution map according to the variable detection location.

9. The method according to claim 8, characterized in that The calculation of the comprehensive risk assessment results of the overall production is specifically: R 总 =W RAI ·RAI+W FRI ·FRI+W ρ ·ρ RAI,FRI (t)+W S ·∑ i,j S ij ; Among them, R 总 is the comprehensive risk score, W ρ , W S , W RAI , W FRI They are the correlation coefficient between the risk accumulation index and the sudden risk index, the linkage strength, and the weight coefficient of the risk accumulation index and the sudden risk index.

10. Sodium dithionite production risk assessment system based on multi-source data fusion, characterized in that: The system comprises: The production process data collection module is used to collect production process data, including environmental data, production process data, equipment data, and operation behavior data, and add time series labels to the data and pre-process the collected data; The key variable linkage analysis module is used to build a multivariate causal association model based on the characteristics of sodium dithionite production, analyze the causal relationship between humidity, dust concentration, equipment status and reaction temperature, and capture the linkage effect of key variables; Real-time risk detection module, which is used to combine production process data, establish detection models, monitor and classify the current risk status in real time, including normal status, chronic cumulative risk and sudden risk, identify the risk change trend, and obtain the risk accumulation index and sudden risk index; Dynamic risk trigger analysis module, which is used to set dynamic risk trigger thresholds based on causal relationships, analyze the correlation between the risk accumulation index and the sudden risk index and the dynamic risk trigger thresholds, and dynamically update the monitoring strategy; The comprehensive risk score calculation module is used to analyze the relationship between the risk accumulation index and the sudden risk index, and calculate the comprehensive risk score of the production line based on the linkage relationship strength, risk accumulation index and sudden risk index, and visually mark the risk situation and risk degree.

Images