Initiating explosive device processing safety assessment and early warning method based on Noisy-OR Bayesian network

Abstract

The invention discloses an initiating explosive device processing safety assessment and early warning method based on a Noisy-OR Bayesian network. According to the method, a Noisy-OR model and a Bayesian network are combined, and an initiating explosive device processing process safety evaluation model is constructed. According to the method, a Noisy-OR model can be used for carrying out quantitative processing on expert evaluation information endowed with weight, a corresponding CPT (conditional probability table) is generated, and the risk probability of the initiating explosive device processing process is generated through a Bayesian network by combining real-time monitoring data of a dangerous point and simulation data of an initiating explosive device processing finite element model through a multi-source information fusion method. According to the method, subjectivity of traditional initiating explosive device processing safety evaluation and singleness of a safety monitoring method are avoided, the Noisy-OR model and the Bayesian network are fused and applied to the field of initiating explosive device processing safety monitoring, the problems that safety evaluation is not objective and not real-time, and the monitoring method is local are solved, and the safety of initiating explosive device processing is improved. The method has a practical guiding significance for monitoring the actual processing safety of the initiating explosive device.

Description

Technical Field

[0001] The present invention belongs to the technical field of pyrotechnics processing, and in particular relates to an application method for safety assessment and early warning of pyrotechnics processing based on Noisy-OR Bayesian network. Background Art

[0002] In the context of the rapid development and iteration of the current military manufacturing field, due to the high proportion of manual operation, low efficiency and high reliance on the experience level of operators in traditional pyrotechnic charging technology, the industry is difficult to transform and upgrade in the direction of intelligence, digitalization and unmanned operation. Therefore, it is necessary to introduce new charging processing technology. However, pyrotechnics have the characteristics of high sensitivity, high energy density and high danger, and are extremely sensitive to external stimuli. In actual process production, even small deviations in process parameters and abnormal equipment operation will lead to safety accidents such as combustion and explosion, which will cause catastrophic consequences, including casualties, equipment damage, production line paralysis and other major losses. Therefore, the intelligent development of pyrotechnic processing technology must be based on strict safety guarantees.

[0003] The current field of pyrotechnics processing faces problems such as incomplete safety monitoring methods and inconsistent safety assessment standards. In particular, traditional single-point monitoring systems primarily focus on individual hazard sources, such as temperature and pressure data, and fail to fully consider the coupling mechanisms between multiple hazard sources during the processing process, nor the overall safety of the system. Furthermore, discrete safety assessment indicators lack systematic linkage analysis. Therefore, to improve the safety monitoring level of pyrotechnics processing, this paper introduces the Noisy-OR Bayesian network model, combines finite element simulation analysis with real-time sensor monitoring data, and constructs a multi-source information fusion safety risk assessment system to obtain the overall risk probability of the pyrotechnics processing process, thereby accurately quantifying the overall risk probability of the processing process.

[0004] Introducing the Noisy-OR Bayesian network model into pyrotechnic processing safety monitoring is a critical step in intelligent monitoring. In industrial engineering, the Noisy-OR Bayesian network model is widely used for fault diagnosis and predictive maintenance. It maintains high parameter efficiency and strong adaptability in addressing practical engineering problems. Applying this model to the pyrotechnic processing sector transforms fuzzy expert assessments, discrete monitoring data, and finite element simulation data into continuous risk probabilities, providing a quantitative safety decision-making foundation for the intelligent development of pyrotechnic processing and possessing direct engineering application value. Summary of the Invention

[0005] In response to the above technical problems, the present invention provides an application method for the safety assessment and early warning of pyrotechnic processing based on the Noisy-OR Bayesian network, aiming to combine the Noisy-OR Bayesian network, finite element simulation and real-time monitoring data to achieve the purpose of accurate real-time monitoring of the safety of pyrotechnic processing.

[0006] 1. The present invention and technical solution is: an application method for safety assessment and early warning of explosive device processing based on Noisy-OR Bayesian network, characterized by comprising the following steps:

[0007] A. Determine the number of hazardous sources in the pyrotechnics processing process as n, invite experts to evaluate the hazardous sources, and use the expert evaluation parameters as the input parameters X of the Noisy-OR model. t , the obtained risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model;

[0008] B. Establish a finite element analysis model for pyrotechnics processing and select material parameters to obtain data on the change in physical properties during pyrotechnics processing;

[0009] C. During the pyrotechnics processing, sensors are connected to obtain real-time monitoring data for detectable hazardous sources, such as temperature, pressure and other parameters;

[0010] D. Integrate the processing simulation data in step B and the real-time monitoring data in step C as the Bayesian network input parameter X through the multi-source information fusion method, combine it with the conditional probability table P calculated in step A, and use the final processing risk probability calculated by the Bayesian network as the output parameter Y. A , get the pyrotechnic processing safety monitoring results, the specific steps are:

[0011] D1. Process the finite element simulation data obtained in step B and the sensor detection data obtained in step C, calculate the difference between the corresponding data with the same attribute, set a reasonable threshold based on the working conditions, and convert the difference and threshold into occurrence probability;

[0012] D2. Based on the conditional probability table, use the joint probability formula to calculate the occurrence probability N of each node combination. The calculation formula is:

[0013]

[0014] Where xi∈{0,1}. Calculate the final processing risk probability by multiplying the joint probability and conditional probability of each combination in the conditional probability table and adding all the results to get the final processing risk probability Y A , the formula is:

[0015]

[0016] Where Mi is the conditional probability of the node combination, and Ni is the occurrence probability of the node combination.

[0017] D3. Based on the actual pyrotechnics processing situation, the final processing risk probability, that is, the output parameter Y A The probability level is divided and the safety threshold is set. If Y A If the value exceeds the threshold, the production system is deemed unsafe and an alarm or emergency stop operation can be performed.

[0018] 2. The method according to claim 1, wherein, in step A, the hazard sources in the pyrotechnics processing process are determined, experts are invited to evaluate the hazard sources, the expert evaluation parameters are used as the input parameters Xt of the Noisy-OR model, and the obtained risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model.

[0019] A1. Determine information on the material properties and processing methods of pyrotechnics. Classify the hazards of pyrotechnic processing by researching literature and querying historical accident databases. Establish a hierarchical safety evaluation system, identify a set of process safety influencing factors, and construct a safety evaluation model.

[0020] A2. Invite university teachers with relevant research interests and enterprise safety personnel to assign corresponding weights based on their expertise. The experts then conduct risk assessments on hazardous sources based on empirical reference tables. The expert reviews are converted into fuzzy probability scores, and the Noisy-OR model is used to calculate the conditional probability table P. The formula for calculating conditional probability using the Noisy-OR model is:

[0021]

[0022] In the formula, Y∈{0,1} indicates whether the result occurs (1 means the result occurs, 0 means the result does not occur), Xi∈{0,1} indicates whether the cause occurs (1 means the cause occurs, 0 means the cause does not occur), and pi is the independent probability that Y=1 when Xi=1 (i.e., the probability that Xi activates Y alone). The conditional probability should be 2 n , corresponding to the conditional probabilities of different nodes occurring.

[0023] 3. The method according to claim 1, wherein in step B, the specific steps of establishing a finite element analysis model for pyrotechnics processing and selecting simulation parameters are as follows:

[0024] B1. Establish a three-dimensional structural model of pyrotechnics processing based on the actual processing site;

[0025] B2. In the finite element simulation analysis software, set the material properties of the pyrotechnic processing model. The basic physical properties of the material include density, Young's modulus, Poisson's ratio, yield strength, and ultimate strength. If you want to pay attention to temperature changes during processing, you need to set the material's thermodynamic properties, including thermal conductivity, thermal expansion coefficient, and specific heat.

[0026] B3. Mesh the pyrotechnics processing model and set the initial conditions, boundary conditions, and analysis settings. The specific meshing type should be based on the structure of the processing model, and an appropriate tetrahedral or hexahedral mesh should be selected. Mesh key locations should be subdivided. Initial and boundary conditions include initial ambient temperature, constraint force conditions, velocity inlet / outlet, etc.

[0027] B4. Determine the model's structural motion mode and parameters based on the specific processing equipment and process. Structural motion modes include extrusion, rolling, screw extrusion, etc. Motion parameters include displacement speed, frequency, rotational angular velocity, etc., which are determined based on different structural motion modes.

[0028] B5. Obtain finite element simulation results, process the results, and store them in the pyrotechnics processing database, including the numerical change data of key physical properties during the pyrotechnics processing.

[0029] 4. The method according to claim 1, wherein in step C, for a hazardous source that can be detected, such as temperature, pressure, or other parameters, connecting a sensor to obtain real-time monitoring data is performed in the following specific steps:

[0030] C1. Determine the physical properties that can be detected by sensors during the pyrotechnics processing process, such as temperature, pressure, and other parameters. Select the corresponding physical sensors and add the appropriate number of sensors based on the specific pyrotechnics processing model and working conditions.

[0031] C2. During the pyrotechnics processing, the data detected by the sensor in real time is analyzed through the serial port program to obtain specific values ​​and store them in the database.

[0032] The present invention provides a method for monitoring the safety of explosive device processing based on a Noisy-OR Bayesian network. This method, for the first time, combines Noisy-OR Bayesian networks, finite element simulation, and real-time sensor monitoring data in the explosive device field. This method addresses the challenges of pyrotechnic process safety monitoring, which suffer from weak objectivity, poor real-time performance, and high locality, meeting the high efficiency and safety requirements of enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1This is a flow chart of an application method of the Noisy-OR Bayesian network in safety assessment and early warning of explosives processing in the present invention.

[0034] Figure 2 Schematic diagram of the Bayesian network in the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose and technical solution of the present invention clearer, the following is a detailed description with reference to specific examples. It should be understood that the specific implementation examples described herein are only used to explain the present invention, but are not limited to the present invention.

[0036] like Figure 1 The figure shows a flow chart of a method for applying the Noisy-OR Bayesian network to the safety assessment and early warning of explosive device processing. The method for applying the Noisy-OR Bayesian network to the safety assessment and early warning of explosive device processing includes the following steps:

[0037] A. Identify the hazardous sources in the pyrotechnics processing process and invite experts to evaluate the hazardous sources. The expert evaluation parameters are used as the input parameters Xt of the Noisy-OR model, and the resulting risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model.

[0038] B. Establish a finite element analysis model for pyrotechnics processing and select material parameters to obtain data on the change in physical properties during pyrotechnics processing;

[0039] C. During the pyrotechnics processing, sensors are connected to obtain real-time monitoring data for detectable hazardous sources, such as temperature, pressure and other parameters;

[0040] D. Integrate the processing simulation data in step B and the real-time monitoring data in step C as the Bayesian network input parameter X through the multi-source information fusion method, combine it with the conditional probability table M calculated in step A, and use the final processing risk probability calculated by the Bayesian network as the output parameter Y. A , obtain the safety monitoring results of pyrotechnic processing;

[0041] In step A, the hazard sources in the pyrotechnics processing process are determined, and experts are invited to evaluate the hazard sources. The expert evaluation parameters are used as the input parameters Xt of the Noisy-OR model, and the obtained risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model. The steps are as follows:

[0042] A1. Determine information about the material properties and processing methods of pyrotechnics. Classify the hazardous sources in the pyrotechnic processing process by researching literature and querying historical accident databases. Thirty-one hazardous sources were identified. A hierarchical safety assessment system was established, with six groups ([4, 4, 5, 6, 6, 6]) forming the intermediate nodes. These six nodes constitute the final leaf nodes, resulting in a set of process safety influencing factors. This led to the construction of a safety assessment model.

[0043] A2. Set the expert language evaluation to correspond to fuzzy numbers. The expert language evaluation is divided into [VL, L, RL, M, RH, H, VH], which correspond to triangular fuzzy numbers [(0, 0, 0.1), (0, 0.1, 0.3), (0.1, 0.3, 0.5), (0.3, 0.5, 0.7), (0.5, 0.7, 0.9), (0.7, 0.9, 1), (0.9, 1, 1)] respectively. Invite 4 teachers from relevant research directions of universities and enterprise safety staff, and assign corresponding weights according to their expert level, which are [0.30, 0.25, 0.25, 0.20] respectively. The 4 experts conduct risk assessment on 31 hazardous sources based on the experience reference table, convert the expert review into fuzzy possibility scores, and store all fuzzy possibility scores in a table as the input parameter Xt of the Noisy-OR model. Take the conditional probability of 3 hazardous nodes leading to the occurrence of the next node X1 as an example evaluated by 3 experts, as shown in Table 1 below:

[0044] Table 1 Scoring table

[0045]

[0046] The Noisy-OR model is used to calculate the conditional probability table P using the fuzzy likelihood score, as shown in Table 2. The formula for calculating the conditional probability using the Noisy-OR model is:

[0047]

[0048] Where Y∈{0,1} indicates whether the result occurs (1 for the result to occur, 0 for the result not to occur), Xi∈{0,1} indicates whether the cause occurs (1 for the cause to occur, 0 for the cause not to occur), pi is the independent probability that Y=1 when Xi=1 (i.e., the probability that Xi activates Y alone), and the conditional probability should be 2 n , corresponding to the conditional probabilities of different nodes occurring.

[0049] Table 2 Rating table

[0050]

[0051] In step B, the specific steps of establishing a finite element analysis model for pyrotechnics processing and selecting simulation parameters are as follows:

[0052] B1. Build a 3D structural model of the pyrotechnics based on the actual processing site. This includes the 3D dimensions, with an overall diameter of 20mm and a height of 90mm.

[0053] B2. In the finite element simulation analysis software, set the parameters of the material properties of the pyrotechnic processing model. The material property settings are shown in Table 3 below.

[0054] Table 3 Attribute table

[0055] density Young's modulus Poisson's ratio Yield strength <![CDATA[1.84g / cm 3 ]]> 7Gpa 0.30 15MPa Ultimate strength Specific heat Thermal conductivity Coefficient of thermal expansion 30MPa 1.1J / (g*K) 0.3W / (m*K) <![CDATA[70*10 -6 K -1 ]]>

[0056] In step (3), the meshing type is determined to be hexahedral unit division, and the mesh size is 2 mm; the initial conditions are set to the initial ambient temperature of 22 ° C and the heat convection heat transfer coefficient of 80 W / m 2 ℃, the constraint is of Bonded type.

[0057] B3. Set the mesh type to hexahedral elements and the mesh size to 0.7 mm. Set the initial ambient temperature to 25°C, constrain the upper and lower ends of the model, and set the velocity inlet to 0.05 mm / s.

[0058] B4. Determine the model structure motion mode as extrusion motion, with a frequency of 1.5 Hz and a displacement of 10 mm.

[0059] B5. Obtain finite element simulation results, process the results, and store them in the pyrotechnics processing database, including the numerical change data of key physical properties during the pyrotechnics processing.

[0060] In step C, for the hazardous source that can be detected, such as temperature, pressure and other parameters, the specific steps of connecting sensors to obtain real-time monitoring data are as follows:

[0061] C1. Determine the physical properties that can be detected by sensors during the pyrotechnics processing process, such as temperature, pressure, and other parameters. Select the corresponding physical sensors and add the appropriate number of sensors based on the specific pyrotechnics processing model and working conditions.

[0062] C2. During the pyrotechnics processing, the data detected by the sensor in real time is analyzed through the serial port program to obtain specific values ​​and store them in the database.

[0063] D. Using the processing simulation data in step B and the real-time monitoring data processing results in step C as the Bayesian network input parameter X, combined with the conditional probability table M calculated in step A, and using the final processing risk probability calculated by the Bayesian network as the output parameter YA, the pyrotechnics processing safety monitoring results are obtained. The specific steps are as follows:

[0064] D1. Process the finite element simulation data obtained in step B and the sensor detection data obtained in step C, calculate the difference between the corresponding data with the same attribute, set a reasonable threshold based on the working conditions, and convert the difference and threshold into occurrence probability;

[0065] D2. Based on the conditional probability table, use the joint probability formula to calculate the occurrence probability N of each node combination. The calculation formula is:

[0066]

[0067] Where xi∈{0,1}. Calculate the final processing risk probability by multiplying the joint probability and conditional probability of each combination in the conditional probability table and adding all the results to get the final processing risk probability Y A , the formula is:

[0068]

[0069] Where Mi is the conditional probability of the node combination, and Ni is the occurrence probability of the node combination.

[0070] D3. Based on the actual pyrotechnics processing situation, the final processing risk probability, that is, the output parameter Y A The probability level is divided into 5 levels, the higher the level, the less safe it is, and the safety threshold level is set to 4. If Y A If the value of is greater than or equal to level 4, the processing system is deemed unsafe and an alarm or emergency stop operation can be performed. A It is 0.83%, which is at level 3, so the processing system is considered safe.

[0071] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A safety assessment and early warning method for explosives processing based on Noisy-OR Bayesian network, characterized by: The following steps are involved: A. Determine the number of hazardous sources in the pyrotechnics processing process as n, invite experts to evaluate the hazardous sources, and use the expert evaluation parameters as the input parameters X of the Noisy-OR model. t , the obtained risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model; B. Establish a finite element analysis model for pyrotechnics processing and select material parameters to obtain data on the change in physical properties during pyrotechnics processing; C. During the pyrotechnics processing, sensors are connected to obtain real-time monitoring data for detectable hazardous sources, such as temperature, pressure and other parameters; D. Integrate the processing simulation data in step B and the real-time monitoring data in step C as the Bayesian network input parameter X through a multi-source information fusion method. Combined with the conditional probability table P calculated in step A, the final processing risk probability calculated by the Bayesian network is used as the output parameter YA to obtain the pyrotechnics processing safety monitoring results. The specific steps are as follows: D1. Process the finite element simulation data obtained in step B and the sensor detection data obtained in step C, calculate the difference between the corresponding data with the same attribute, set a reasonable threshold based on the working conditions, and convert the difference and threshold into occurrence probability; D2. Based on the conditional probability table, use the joint probability formula to calculate the occurrence probability N of each node combination. The calculation formula is: Where xi∈{0,1}, calculate the final processing risk probability, and multiply the joint probability and conditional probability of each combination in the conditional probability table and add all the results to get the final processing risk probability Y A , the formula is: Where Mi is the conditional probability of the node combination, Ni is the occurrence probability of the node combination; D3. Based on the actual pyrotechnics processing situation, the final processing risk probability, that is, the output parameter Y A The probability level is divided and the safety threshold is set. If Y A If the value exceeds the threshold, the production system is deemed unsafe and an alarm or emergency stop operation can be performed.

2. The method according to claim 1, wherein: In step A, the hazard sources in the pyrotechnics processing process are determined, and experts are invited to evaluate the hazard sources. The expert evaluation parameters are used as the input parameters Xt of the Noisy-OR model, and the obtained risk transfer conditional probability table is used as the output parameter P of the Noisy-OR model. The steps are as follows: A1. Determine information on the material properties and processing methods of pyrotechnics. Classify the hazards of pyrotechnic processing by researching literature and querying historical accident databases. Establish a hierarchical safety evaluation system, identify a set of process safety influencing factors, and construct a safety evaluation model. A2. Invite university teachers with relevant research interests and enterprise safety personnel to assign corresponding weights based on their expertise. The experts then conduct risk assessments on hazardous sources based on empirical reference tables. The expert reviews are converted into fuzzy probability scores, and the Noisy-OR model is used to calculate the conditional probability table M. The formula for calculating conditional probabilities using the Noisy-OR model is: Where Y∈{0,1} indicates whether the result occurs (1 for the result to occur, 0 for the result not to occur), Xi∈{0,1} indicates whether the cause occurs (1 for the cause to occur, 0 for the cause not to occur), pi is the independent probability that Y=1 when Xi=1 (i.e., the probability that Xi activates Y alone), and the conditional probability should be 2 n , corresponding to the conditional probabilities of different nodes occurring.

3. The method according to claim 1, wherein: In step B, the specific steps of establishing a finite element analysis model for pyrotechnics processing and selecting simulation parameters are as follows: B1. Establish a three-dimensional structural model of pyrotechnics processing based on the actual processing site; B2. In the finite element simulation analysis software, set the material properties of the pyrotechnic processing model. The basic physical properties of the material include density, Young's modulus, Poisson's ratio, yield strength, and ultimate strength. If you want to pay attention to temperature changes during processing, you need to set the material's thermodynamic properties, including thermal conductivity, thermal expansion coefficient, and specific heat. B3. Mesh the pyrotechnics processing model and set the initial conditions, boundary conditions, and analysis settings. The specific meshing type should be based on the structure of the processing model, and an appropriate tetrahedral or hexahedral mesh should be selected. Mesh key locations should be subdivided. Initial and boundary conditions include initial ambient temperature, constraint force conditions, velocity inlet / outlet, etc. B4. Determine the model's structural motion mode and parameters based on the specific processing equipment and process. Structural motion modes include extrusion, rolling, screw extrusion, etc. Motion parameters include displacement speed, frequency, rotational angular velocity, etc., which are determined based on different structural motion modes. B5. Obtain finite element simulation results, process the results, and store them in the pyrotechnics processing database, including the numerical change data of key physical properties during the pyrotechnics processing.

4. The method according to claim 1, wherein: In step C, for the hazardous source that can be detected, such as temperature, pressure and other parameters, the specific steps of connecting sensors to obtain real-time monitoring data are as follows: C1. Determine the physical properties of the pyrotechnics process that can be detected by sensors, such as temperature, pressure, and other parameters, select the corresponding physical sensors, and add the appropriate number of sensors based on the specific pyrotechnics processing model and working conditions; C2. During the pyrotechnics processing, the data detected by the sensor in real time is analyzed through the serial port program to obtain specific values ​​and store them in the database.

Images