Emergency scheduling optimization method and system for chemical safety production

Abstract

The invention discloses an emergency scheduling optimization method and system for chemical safety production, relates to the technical field of industrial production scheduling, and solves the problems that a traditional emergency scheduling method provides a scheduling processing scheme excessively depending on field expert experience, and chemical products which are produced uninterruptedly and safely cannot be dispatched all the time and cannot be dispatched continuously and safely. People with rich experiences have significant potential risks in nursing. Meanwhile, the problems that monitoring, early warning and scheduling schemes are only provided at a device level, systematic thinking is lacked, and balance of a production system cannot be achieved only by changing device parameters are solved; according to the method, a preset event set is established through historical cases, the preset event set comprises preset accidents and preset measures, the preset accidents comprise preset abnormal working conditions, each case is analyzed, the historical cases are classified, a batch of cases with the best accident coping effect are selected as a standard, and the cases are analyzed; relations among indexes such as accident types, processing measures, processing time, prepared material consumption and accident loss are established.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial production scheduling, and in particular to an emergency scheduling optimization method and system for safe production of chemical industry. Background Art

[0002] In modern society, the application of chemicals has penetrated into various fields, from daily necessities to high-tech products, all of which are inseparable from the support of chemical products. However, the potential dangers in the chemical production process cannot be ignored; therefore, chemical production safety has become an important issue of common concern within the industry, government supervision and the whole society; the emergency dispatch optimization of chemical production safety is to be able to quickly and effectively allocate resources and dispatch processing when an emergency occurs, so as to minimize the impact of the accident; this method usually includes multiple key steps, from monitoring of abnormal conditions to calling and executing emergency plans, and then to optimization calculation and adjustment, forming a closed-loop emergency management system;

[0003] In-depth analysis of emergency management work shows that its essence is to achieve effective risk control by reducing emergency response time; chemical production safety accidents are unpredictable and will have a serious impact on production operations once they occur; therefore, in actual work, emergency response measures must be taken within a short period of time to reduce the losses caused by safety accidents and the adverse effects on all aspects of chemical production safety operations;

[0004] Traditional emergency dispatch methods rely too much on the experience of on-site experts to provide dispatch solutions. For chemical products that require uninterrupted safe production, there cannot be experienced people to supervise them at all times, which poses significant potential risks. At the same time, monitoring, early warning and dispatch solutions are only provided at the device level, lacking systematic thinking, and the balance of the production system cannot be achieved by simply changing the device parameters. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an emergency dispatch optimization method and system for chemical production safety, which solves the problem that the traditional emergency dispatch method relies too much on the experience of on-site experts to provide dispatch processing solutions. For chemical products with uninterrupted safe production, it is impossible to have experienced people to watch over them all the time, which poses a major potential risk. At the same time, it only provides monitoring, early warning and dispatch solutions at the device level, lacks systematic thinking, and cannot achieve the balance of the production system by only changing the device parameters.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an emergency dispatch optimization method for chemical production safety, comprising the following steps:

[0007] S1. Establish a preset event set through historical cases, including preset accidents and preset measures, where preset accidents include preset abnormal conditions, and analyze each case:

[0008] S2. Collect dynamic data of the production line, establish an emergency dispatch model, set up a visual flow chart, and systematically record and analyze the key import and export flows of each device zi for the entire production line;

[0009] S3. Use the qualified event set to calibrate and optimize the emergency dispatch model, generate an emergency dispatch optimization model, optimize the preset measures in the unqualified event set, perform data analysis on the preset measures in the extremely unqualified event set, and regenerate the preset measures.

[0010] Furthermore, the analysis of abnormal operating conditions in historical cases includes the following steps:

[0011] For each historical case's accident type, handling measures, handling time, reserve material consumption and accident loss, analyze the relationship between the case's handling time and the accident loss value based on the accident type. The analysis steps are as follows:

[0012]

[0013] Where C is the treatment coefficient, is a constant, L is the consumption of prepared materials, T is the processing time, S is the risk factor of the accident category, and A is the accident loss. Based on the processing coefficient C, historical cases are divided into qualified event set, unqualified event set and extremely unqualified event set.

[0014] Furthermore, the division of historical cases into a qualified event set, an unqualified event set and an extremely unqualified event set based on the processing coefficient C comprises the following steps:

[0015] Define the processing coefficient c1 of the qualified event set and the processing coefficient c2 of the unqualified event set, and mark the preset measures in the qualified event set as the recommended measures for the event, the preset measures in the unqualified event set as the measures to be optimized for the event, and the preset measures in the extremely unqualified event set as the measures to be replaced for the event;

[0016] According to the processing coefficient c1 and the processing coefficient c2, the preset event set is divided into three event sets: qualified, unqualified and extremely unqualified;

[0017] The division steps are as follows: counting the processing coefficient of each historical case, and classifying the historical cases with processing coefficients greater than the processing coefficient c1 into the qualified event set;

[0018] The historical cases with treatment coefficient less than or equal to treatment coefficient c1 and greater than treatment coefficient c2 are classified into the unqualified event set;

[0019] The historical cases with treatment coefficient less than or equal to treatment coefficient c2 are classified into the extremely unqualified event set.

[0020] Furthermore, the production line dynamic data collection includes the following steps:

[0021] The production line dynamic data collection steps are as follows: according to the production order, by numbering each device, recorded as zi, and setting a sensor group on each device on the production line, the inflow and outflow of materials in each device zi, and the real-time collection of the state of the reaction chamber, wherein i is the serial number of each device;

[0022] The sensor group includes a temperature sensor, a humidity sensor, a carbon dioxide sensor and a pressure sensor, and is numbered in sequence and denoted as j.

[0023] Furthermore, the establishment of the emergency dispatch model includes the following steps:

[0024] Through deep learning algorithms, the normal operation data of the production line is called for training to analyze the input and output of each device zi in the entire production line and the fluctuation of key indicators;

[0025] The key streams of each device are range-marked, where the inlet key stream is denoted as Lk, the outlet key stream is denoted as lk, and k is the stream sequence number; the temperature is denoted as t; the humidity is denoted as s; the carbon dioxide concentration is denoted as c; the pressure value is denoted as y; the fluctuation ranges are set respectively, and the setting method is to extract the ranges of indicators calculated by the deep learning algorithm; record the various indicators of each device zi, and form a unique range interval set for each device zi.

[0026] Furthermore, the drawing of the visual flowchart includes the following steps:

[0027] Define three alarm states in the visual flow chart: normal operation alarm, abnormal alarm and severe abnormal alarm;

[0028] The normal operation alarm indicates that the temperature, humidity, carbon dioxide concentration and pressure values ​​of the key inflow and outflow streams in the device zi are all within the range interval set of the device zi, and any indicator does not have a continuous t1 period of time, exceeding the preset range interval; it indicates the production status of each device in chemical safety production.

[0029] Abnormal alarm, indicating that at least one of the key inflow and outflow streams, humidity and carbon dioxide concentration in device zi is outside the range interval set of the device zi, and the continuous time is greater than the time period t1;

[0030] Severe abnormal alarm, indicating that at least one of the temperature and pressure values ​​in device zi is outside the range interval set of device zi, and the duration is greater than t2;

[0031] The priorities of the three alarm modes are as follows: severe abnormal alarm is greater than abnormal alarm and normal operation alarm, and abnormal alarm is greater than normal operation alarm;

[0032] According to the device zi number, each device zi is connected with the device zi-1 and device zi+1 before and after the chemical safety production line by arrows, and the direction of the arrow is the direction of the stream flow; three warning lights are set next to each device zi, among which the green warning light represents the normal operation alarm, the yellow warning light represents the abnormal alarm, and the red warning light represents the severe abnormal alarm.

[0033] Furthermore, the method of calibrating and optimizing the emergency dispatch model using the qualified event set to generate the emergency dispatch optimization model includes the following steps:

[0034] Taking each device zi as a unit, select the cases that occurred in the device zi from the qualified event set to generate a case library; select the m single indicator data with the closest time distance in the case library, such as temperature, and calculate the mean of the m temperatures and variance t σ ; Take the mean is the midpoint, variance t σ For the radius, generate the temperature optimization interval Replace the temperature fluctuation interval with the temperature optimization interval;

[0035] Similarly, for the humidity s, carbon dioxide concentration c, inlet key stream Lk, outlet key stream lk and pressure value y in the case library, their respective optimization ranges are calculated; among them, m is pre-set by the staff based on experience.

[0036] Furthermore, the optimization of preset measures in the set of unqualified events, data analysis of preset measures in the set of extremely unqualified events, and regeneration of preset measures include the following steps:

[0037] Define W as the dynamic balance coefficient inside the device, and the value range is [w1, w2], which means the fluctuation range of W value when all indicators inside the device are within the optimization range. Analyze the dynamic balance formula:

[0038] W=α0+α1*t+α2*s+α3*c+α4*Lk+α5*y+α6*t*s+α7*t*c+α8*s*c+α9*Lk*y+∈;

[0039] Among them, w1 and w2 are the upper and lower limits of the dynamic balance coefficient value range, α0, α2, α3, α4, α5, α6, α7, α8, and α9 are model parameters, and ∈ is the error term, which are all calculated by the staff for each device zi through the deep learning algorithm;

[0040] For the measures to be optimized and replaced, the dynamic balance formula is used to constrain the index weights in the device zi, such as temperature ti; humidity si; carbon dioxide concentration ci; pressure value yi and the inlet key stream denoted as Lki and the outlet key stream denoted as lki. The constraint steps are as follows;

[0041] Define stable accidents and sudden accidents. A stable accident is when the rate of change of abnormal indicators in the accident is greater than the normal rate v1p and less than or equal to the lower limit rate v2p; a sudden accident is when the rate of change of abnormal indicators in the accident is greater than the rate v3p; where p is the indicator type, and the values ​​of v1p, v2p, and v3p are all calculated by the staff for each device zi through a deep learning algorithm;

[0042] For stable accidents, abnormal alarms are activated, and dynamic balance analysis is performed on abnormal indicators. By adjusting other indicators with control space, the dynamic balance coefficient of the device is forced to return to the dynamic balance value range;

[0043] In case of sudden changes, a severe abnormal alarm is activated, the inlet flow of the device is cut off, and the production device of the upstream flow is automatically shut down;

[0044] For the device zi that has not experienced any accidents, the emergency dispatch optimization model calculates the optimization range of each indicator of the device zi based on the chemical production data and the internal dynamic balance formula of the device zi;

[0045] And according to each device zi, a preset measure library is generated, including recommended measures, and suggested measures given by the emergency dispatch optimization model for measures to be optimized and measures to be replaced.

[0046] The invention discloses an emergency dispatch optimization system for safe production, which specifically adopts an emergency dispatch optimization method for safe production in chemical industry to realize emergency dispatch for safe production in chemical industry.

[0047] Beneficial Effects

[0048] The present invention provides an emergency dispatch optimization method and system for chemical production safety. Compared with the prior art, it has the following beneficial effects:

[0049] The present invention establishes a preset event set through historical cases, including preset accidents and preset measures, wherein the preset accidents include preset abnormal working conditions, analyzes each case, classifies the historical cases, selects a batch of cases with the best accident response effects as standards, and on this basis, analyzes the relationship between the cases, accident categories, treatment measures, treatment time, consumption of prepared materials, accident losses and other indicators, calculates the treatment coefficient, and classifies all historical cases based on the treatment coefficient; in actual cases, the safety range of chemical production is found, which can more reasonably allocate production resources, improve production efficiency, ensure production continuity, and avoid the huge difference between theoretical analysis and actual conditions;

[0050] Collect dynamic data of the production line, establish an emergency dispatch model, set up a visual flow chart, and systematically record and analyze the key import and export flows of each device zi for the entire production line; change the fluctuation range of each indicator according to the current usage of each device. Reasonable adjustment of the fluctuation range can more accurately identify potential risk points, so as to take corresponding risk control measures; be more sensitive to production changes, be able to respond quickly, reduce the impact of adverse production changes on the device, and set a reasonable loss limit to prevent substantial losses; by pre-setting parameters and dynamic analysis of indicators specific to each device, it solves the problem of excessive reliance on workers' experience when abnormal working conditions occur, or the dilemma of no precedents for unrecorded abnormal working conditions;

[0051] The present invention uses the qualified event set to calibrate and optimize the emergency dispatch model, generates an emergency dispatch optimization model, optimizes the preset measures in the unqualified event set, performs data analysis on the preset measures in the extremely unqualified event set, and regenerates the preset measures;

[0052] Through continuous updating, the emergency dispatch optimization model optimized by specific cases re-analyzes the measures to be optimized and replaced, and gives better solutions to adapt to the ever-changing production process and emerging safety risks; this includes the ability to respond quickly to newly identified risk factors, and uses real-time data and historical data analysis to improve the accuracy of model predictions and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flow chart of the method of the present invention;

[0054] Figure 2 This is a system block diagram of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Embodiment 1

[0057] See also Figure 1 The present application provides an emergency dispatch optimization method for chemical production safety, comprising the following steps:

[0058] S1. Establish a preset event set through historical cases, including preset accidents and preset measures, where the preset accidents include preset abnormal conditions. Analyze each case, including the following steps:

[0059] For each historical case's accident type, handling measures, handling time, reserve material consumption and accident loss, analyze the relationship between the case's handling time and the accident loss value based on the accident type. The analysis steps are as follows:

[0060]

[0061] Where C is the treatment coefficient, is a constant, L is the consumption of prepared materials, T is the processing time, S is the risk factor of the accident category, and A is the accident loss; based on the processing coefficient C, historical cases are divided into qualified event sets, unqualified event sets, and extremely unqualified event sets, including the following steps:

[0062] Define the processing coefficient c1 of the qualified event set and the processing coefficient c2 of the unqualified event set, and mark the preset measures in the qualified event set as the recommended measures for the event, the preset measures in the unqualified event set as the measures to be optimized for the event, and the preset measures in the extremely unqualified event set as the measures to be replaced for the event;

[0063] According to the processing coefficient c1 and the processing coefficient c2, the preset event set is divided into three event sets: qualified, unqualified and extremely unqualified;

[0064] The division steps are as follows: counting the processing coefficient of each historical case, and classifying the historical cases with processing coefficients greater than the processing coefficient c1 into the qualified event set;

[0065] The historical cases with treatment coefficient less than or equal to treatment coefficient c1 and greater than treatment coefficient c2 are classified into the unqualified event set;

[0066] The historical cases with treatment coefficient less than or equal to treatment coefficient c2 are classified into the extremely unqualified event set.

[0067] In actual cases, finding the safe range of chemical production can more reasonably allocate production resources, improve production efficiency, ensure production continuity, and avoid the huge gap between theoretical analysis and actual conditions.

[0068] S2. Collect dynamic data of the production line, establish an emergency dispatch model, set up a visual flow chart, and systematically record and analyze the key import and export flows of each device zi for the entire production line;

[0069] The production line dynamic data collection steps are as follows: according to the production order, each device is numbered as zi, and a sensor group is set on each device on the production line to collect the inflow and outflow of materials in each device zi and the state of the reaction chamber in real time.

[0070] The sensor group includes a temperature sensor, a humidity sensor, a carbon dioxide sensor and a pressure sensor, and is numbered in sequence and denoted as j;

[0071] The temperature sensor is used to detect whether the device zi has caught fire or other accidents; the humidity sensor is used to monitor and control the environmental humidity of the production process in the reaction chamber to ensure production quality and safety; the carbon dioxide sensor is used to detect and control the ventilation conditions in the reaction chamber of the device zi; the pressure sensor is used to detect the pressure changes in the reaction chamber of the device zi, and is also used to check whether the inflow and outflow of materials are aligned and whether the reaction device zi has leakage;

[0072] Where i is the serial number of each device; for example, z23 represents the environmental humidity data of the second device;

[0073] The establishment of the emergency dispatch model comprises the following steps:

[0074] Through deep learning algorithms, the normal operation data of the production line is called for training to analyze the input and output of each device zi in the entire production line and the fluctuation of key indicators;

[0075] The key streams of each device are range-marked, where the inlet key stream is denoted as Lk, the outlet key stream is denoted as lk, and k is the stream sequence number; the temperature is denoted as t; the humidity is denoted as s; the carbon dioxide concentration is denoted as c; the pressure value is denoted as y; the fluctuation ranges are set respectively, and the setting method is to extract the range of each indicator calculated by the deep learning algorithm; record each indicator of each device zi, and form a unique range interval set for each device zi;

[0076] Draw a visual flow chart; define three alarm states in the visual flow chart, normal operation alarm, abnormal alarm and severe abnormal alarm;

[0077] The normal operation alarm indicates that the temperature, humidity, carbon dioxide concentration and pressure values ​​of the key inflow and outflow streams in the device zi are all within the range interval set of the device zi, and any indicator does not have a continuous t1 period of time, exceeding the preset range interval; it indicates the production status of each device in chemical safety production.

[0078] Abnormal alarm, indicating that at least one of the key inflow and outflow streams, humidity and carbon dioxide concentration in device zi is outside the range interval set of the device zi, and the continuous time is greater than the time period t1;

[0079] Severe abnormal alarm, indicating that at least one of the temperature and pressure values ​​in device zi is outside the range interval set of device zi, and the duration is greater than t2;

[0080] The priorities of the three alarm modes are as follows: severe abnormal alarm is greater than abnormal alarm and normal operation alarm, and abnormal alarm is greater than normal operation alarm;

[0081] Among them, the values ​​of time periods t1 and t2 are set by staff based on experience;

[0082] Draw a visual flow chart, including the following steps: according to the device zi number, each device zi is connected with the device zi-1 and device zi+1 before and after the device zi in the chemical safety production line by arrows, and the direction of the arrow is the direction of the flow stream; three warning lights are set next to each device zi, among which the green warning light represents the normal operation alarm, the yellow warning light represents the abnormal alarm, and the red warning light represents the severe abnormal alarm; wherein, the yellow warning light and the red warning light are connected to a buzzer, and when the yellow warning light is always on, the buzzer sounds intermittently; when the red warning light is always on, the buzzer sounds long;

[0083] S3. Use the qualified event set to calibrate and optimize the emergency dispatch model, generate an emergency dispatch optimization model, optimize the preset measures in the unqualified event set, perform data analysis on the preset measures in the extremely unqualified event set, and regenerate the preset measures;

[0084] The optimization steps are as follows: taking each device zi as a unit, select the cases that occur in the same device zi from the qualified event set to generate a case library; select the m single indicator data with the closest time distance in the case library, such as temperature, and calculate the mean of the m temperatures and variance t σ ; Take the mean is the midpoint, variance t σ For the radius, generate the temperature optimization interval Replace the temperature fluctuation interval with the temperature optimization interval;

[0085] By limiting the selection to the m single indicator data closest to the time in the case library, it is ensured that the indicator optimization interval fits the actual production, is more sensitive to production changes, and can respond quickly.

[0086] Similarly, for the humidity s, carbon dioxide concentration c, inlet key stream Lk, outlet key stream lk and pressure value y in the case library, their respective optimization ranges are calculated; among them, m is pre-set by the staff based on experience;

[0087] According to the recent use of each device, the fluctuation range of each indicator is changed. Reasonable adjustment of the fluctuation range can more accurately identify potential risk points, so as to take corresponding risk control measures; it is more sensitive to production changes and can respond quickly to reduce the impact of adverse production changes on the device, and can set reasonable loss limits to prevent large losses;

[0088] Define W as the dynamic balance coefficient inside the device, and the value range is [w1, w2], which means the fluctuation range of W value when all indicators inside the device are within the optimization range. Analyze the dynamic balance formula:

[0089] W=α0+α1*t+α2*s+α3*c+α4*Lk+α5*y+α6*t*s+α7*t*c+α8*s*c+α9*Lk*y+∈;

[0090] Among them, w1 and w2 are the upper and lower limits of the dynamic balance coefficient value range, α0, α2, α3, α4, α5, α6, α7, α8, and α9 are model parameters, and ∈ is the error term, which are all calculated by the staff for each device zi through the deep learning algorithm;

[0091] For the measures to be optimized and replaced, the dynamic balance formula is used to constrain the index weights in the device zi, such as temperature ti; humidity si; carbon dioxide concentration ci; pressure value yi and the inlet key stream denoted as Lki and the outlet key stream denoted as lki. The constraint steps are as follows;

[0092] Define stable accidents and sudden accidents. A stable accident is when the rate of change of abnormal indicators in the accident is greater than the normal rate v1p and less than or equal to the lower limit rate v2p; a sudden accident is when the rate of change of abnormal indicators in the accident is greater than the rate v3p; where p is the indicator type, and the values ​​of v1p, v2p, and v3p are all calculated by the staff for each device zi through a deep learning algorithm;

[0093] For stable accidents, abnormal alarms are activated, and dynamic balance analysis is performed on abnormal indicators. By adjusting other indicators with control space, the dynamic balance coefficient of the device is forced to return to the dynamic balance value range;

[0094] For example, if the temperature t increases at a rate between v1p and v2p and exceeds the upper limit of the temperature optimization interval, and the humidity s decreases at a rate between v1p and v2p and is lower than the lower limit of the humidity optimization interval, it is recommended to reduce the internal reaction intensity of the device zi and reduce the input rate of the key inlet stream Lk;

[0095] In case of sudden changes, a severe abnormal alarm is activated, the inlet flow of the device is cut off, and the production device of the upstream flow is automatically shut down;

[0096] For example, if the pressure value y in device zi decreases at a rate exceeding v3p, it is judged that device zi is leaking, and it is recommended to evacuate the crowd, cut off the power supply and notify the relevant departments;

[0097] For the device zi that has not experienced any accidents, the emergency dispatch optimization model calculates the optimization range of each indicator of the device zi based on the chemical production data and the internal dynamic balance formula of the device zi;

[0098] And according to each device zi, a preset measure library is generated, including recommended measures, and suggested measures given by the emergency dispatch optimization model for measures to be optimized and measures to be replaced.

[0099] Through continuous updating, the emergency dispatch optimization model optimized by specific cases re-analyzes the measures to be optimized and replaced, and gives better solutions to adapt to the ever-changing production process and emerging safety risks; this includes the ability to respond quickly to newly identified risk factors, and uses real-time data and historical data analysis to improve the accuracy of model predictions and response speed.

[0100] Embodiment 2

[0101] See also Figure 2 , an emergency dispatch optimization system for chemical production safety, the system uses the above-mentioned emergency dispatch optimization method for chemical production safety to optimize the emergency dispatch of chemical production safety, the system comprises:

[0102] Data management unit: used to process historical cases, establish preset event sets, including preset accidents and preset measures, where preset accidents include preset abnormal conditions, and analyze each case;

[0103] Dynamic management unit: collect dynamic data of production line, establish emergency dispatch model, set up visual flow chart, and systematically record and analyze the key import and export flows of each device for the entire production line;

[0104] Optimization unit: used to calibrate and optimize the emergency dispatch model for the qualified event set, generate the emergency dispatch optimization model, optimize the preset measures in the unqualified event set, perform data analysis on the preset measures in the extremely unqualified event set, and regenerate the preset measures;

[0105] Display unit: used to display the detected indicators of each device on the production line, visualized flow charts, and optimization measures generated by the emergency dispatch optimization model;

[0106] Alarm unit: used to convert the electrical signal of the warning light into the sound signal of the buzzer.

[0107] Embodiment 3

[0108] The specific implementation process of this embodiment includes the entire implementation process of the above two groups of embodiments.

[0109] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0110] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An emergency dispatch optimization method for chemical production safety, characterized in that: The following steps are involved: S1. Establish a preset event set through historical cases, including preset accidents and preset measures, where the preset accidents include preset abnormal conditions. Analyze each case, including the following steps: S2. Collect dynamic data of the production line, establish an emergency dispatch model, set up a visual flow chart, and systematically record and analyze the key import and export flows of each device zi for the entire production line; S3. Use the qualified event set to calibrate and optimize the emergency scheduling model, generate the emergency scheduling optimization model, and optimize the preset measures in the unqualified event set, and conduct data analysis on the preset measures in the extremely unqualified event set, and regenerate the preset measures.

2. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: Analyzing abnormal operating conditions in historical cases includes the following steps: For the accident category, handling measures, handling time, preparatory material consumption and accident loss of each historical case, the connection between the case's handling time and the accident loss value is analyzed according to the accident category, and the analysis steps are; Where C is the treatment coefficient, It is a constant, L is the consumption of reserve materials, T is the processing time, S is the accident category hazard coefficient, and A is the accident loss; based on the processing coefficient C, historical cases are divided into qualified event sets, unqualified event sets and extremely unqualified event sets.

3. The method for optimizing emergency dispatching of chemical production safety according to claim 2, characterized in that: Based on the processing coefficient C, historical cases are divided into qualified event set, unqualified event set and extremely unqualified event set, including the following steps: Define the processing coefficient c1 of the qualified event set and the processing coefficient c2 of the unqualified event set, and mark the preset measures in the qualified event set as the recommended measures for the event, the preset measures in the unqualified event set are recorded as the measures to be optimized for the event, and the preset measures in the extremely unqualified event set are recorded as the measures to be replaced for the event; According to the processing coefficient c1 and the processing coefficient c2, the preset event set is divided into three event sets: qualified, unqualified and extremely unqualified; The division steps are as follows: counting the processing coefficient of each historical case, and classifying the historical cases with processing coefficients greater than the processing coefficient c1 into the qualified event set; The historical cases with treatment coefficient less than or equal to treatment coefficient c1 and greater than treatment coefficient c2 are classified into the unqualified event set; The historical cases with treatment coefficient less than or equal to treatment coefficient c2 are classified into the extremely unqualified event set.

4. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: Dynamic data collection of production line includes the following steps: The dynamic data acquisition step of the production line is: by numbering each device in the production order, denoting it as zi, and setting a sensor group on each device on the production line, the inflow and outflow of materials in each device zi, and real-time collection of the indoor state, where i is the serial number of each device; The sensor group includes a temperature sensor, a humidity sensor, a carbon dioxide sensor and a pressure sensor, and is numbered in sequence and denoted as j.

5. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: Establishing an emergency dispatch model includes the following steps: Through deep learning algorithms, the normal operation data of the production line is called for training to analyze the input and output of each device zi in the entire production line and the fluctuation of key indicators; The key streams of each device are range-marked, where the inlet key stream is denoted as Lk, the outlet key stream is denoted as lk, and k is the stream sequence number; the temperature is denoted as t; the humidity is denoted as s; the carbon dioxide concentration is denoted as c; the pressure value is denoted as y; the fluctuation ranges are set respectively, and the setting method is to extract the ranges of indicators calculated by the deep learning algorithm; record the various indicators of each device zi, and form a unique range interval set for each device zi.

6. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: Draw a visual flowchart, including the following steps: Define three alarm states in the visual flow chart: normal operation alarm, abnormal alarm and severe abnormal alarm; The normal operation alarm indicates the key flow in and out of the device zi. The temperature, humidity, carbon dioxide concentration and pressure values ​​are all within the range set of the device zi. Any indicator does not exist in a continuous t1 time period, exceeding the preset range range; it indicates the production status of each device in chemical production safety. Abnormal alarm, indicating that at least one of the key inflow and outflow streams, humidity and carbon dioxide concentration in device zi is outside the range interval set of the device zi, and the continuous time is greater than the time period t1; Severe abnormal alarm, indicating that at least one of the temperature and pressure values ​​in device zi is outside the range interval set of device zi, and the duration is greater than t2; The priorities of the three alarm modes are as follows: severe abnormal alarm is greater than abnormal alarm and normal operation alarm, and abnormal alarm is greater than normal operation alarm; According to the device number, each device zi is connected with the device zi-1 and device zi+1 in front and behind the chemical safety production line by arrows, and the direction of the arrow is the direction of flow; three alarm lights are arranged next to each device zi, among which the green alarm light represents a normal operation alarm, the yellow alarm light represents an abnormal alarm, and the red alarm light represents a severe abnormal alarm.

7. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: The emergency dispatch model is calibrated and optimized using the qualified event set to generate an emergency dispatch optimization model, including the following steps: In each device, select the same case that occurs in the qualified event set to generate a case library; select m single indicator data closest to the time in the case library, such as temperature, and calculate the mean of m temperatures. and variance tσ; with mean is the midpoint, variance t σ For the radius, generate the temperature optimization interval Replace the temperature fluctuation interval with the temperature optimization interval; Similarly, the respective optimization ranges are calculated for the humidity s, carbon dioxide concentration c, inlet key flow stock Lk, export key flow stock lk and pressure value y in the case library; where m is preset by the staff based on experience.

8. The method for optimizing emergency dispatch for chemical production safety according to claim 1, characterized in that: Optimize the preset measures for the unqualified event set, perform data analysis on the preset measures for the extremely unqualified event set, and regenerate the preset measures, including the following steps: Define W as the dynamic balance coefficient inside the device, and the value range is [w1, w2], which means the fluctuation range of W value when all indicators inside the device are within the optimization range. Analyze the dynamic balance formula: W=α0+α1*t+α2*s+α3*c+α4*Lk+α5*y+α6*t*s+α7*t*c+α8*s*c+α9*Lk*y+∈; Among them, w1 and w2 are the upper and lower limits of the dynamic balance coefficient value range, α0, α2, α3, α4, α5, α6, α7, α8, and α9 are model parameters, and ∈ is the error term, which are all calculated by the staff for each device zi through the deep learning algorithm; For the measures to be optimized and the measures to be replaced, the indicator weights in the constraint device zi, such as temperature ti; humidity si; carbon dioxide concentration ci; pressure value yi and the inlet key flow are recorded as Lki and the export key flow are recorded as Lki, and the constraint steps are; Define a stable accident and a mutation accident. The stable accident is that the abnormal indicator change rate in the accident is greater than the normal rate v1p, less than or equal to the lower limit rate v2p, and the mutation accident is that the abnormal indicator change rate in the accident is greater than the rate v3p; where p is the index type, and the values ​​of v1 p, v2p, and v3p are calculated by the staff for each device through a deep learning algorithm; For stable accidents, an abnormal alarm is activated and a dynamic balance analysis of abnormal indicators is carried out. By adjusting other indicators with regulatory space, the device zi dynamic balance coefficient is prompted to regress to the dynamic balance value range; In case of sudden changes, a severe abnormal alarm is activated, the inlet flow of the device is cut off, and the production device of the upstream flow is automatically shut down; For the device zi that has not experienced any accidents, the emergency dispatch optimization model calculates the optimization range of each indicator of the device zi based on the chemical production data and the internal dynamic balance formula of the device zi; And according to each device zi, a preset measure library is generated, including recommended measures, and suggested measures given by the emergency dispatch optimization model for measures to be optimized and measures to be replaced.

9. An emergency dispatch optimization system for safe production, characterized in that: The system specifically adopts the emergency dispatch optimization method for chemical production safety as described in any one of claims 1 to 8 to realize emergency dispatch for chemical production safety.

Images