Pressure vessel explosion risk prediction method and prediction system

Abstract

The invention discloses a pressure vessel explosion risk prediction method and prediction system, and belongs to the technical field of pressure vessel risk prediction. The prediction method comprises the following steps: step 1, collecting related data of the pressure vessel in real time; 2, the pressure fluctuation rate s, the corrosion speed Rc and the residual strength Sr of the pressure container are calculated; 3, on the basis of the fracture mechanics theory, a critical failure pressure model is built, and the failure probability is obtained; 4, constructing an XGBoost model based on an XGBoost algorithm, inputting characteristic parameters, and carrying out XGBoost operation to obtain a prediction probability; 5, according to the failure probability and the prediction probability, obtaining the mixed risk probability of the pressure vessel after weight distribution; and step 6, according to the mixed risk probability, dividing the pressure vessel into different risk grades. According to the method, dynamic risks can be captured, the method adapts to complex working conditions, the accuracy and reliability of risk early warning can be remarkably improved, and a more scientific early warning scheme is provided for chemical enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical equipment risk prediction, and in particular to a pressure vessel explosion risk prediction method and prediction system. Background Art

[0002] As a key equipment that carries high temperature, high pressure and corrosive media, the safety of pressure vessels is directly related to human life, environmental ecology and corporate assets. The risk of pressure vessel explosion is affected by many factors, such as design and manufacturing, use management, media and environment. Traditional risk prediction methods for pressure vessels mainly rely on regular manual inspections, statistical models based on historical accidents or simplified safety factor methods. The above methods have obvious limitations: for example, manual inspections are highly subjective and have poor real-time performance, making it difficult to capture dynamic risks. Statistical models are not adaptable enough to new materials or complex working conditions. The safety factor method relies too much on empirical thresholds and cannot quantify the failure probability under the coupling of multiple factors. Therefore, there is an urgent need for a dynamic risk assessment method that can integrate structural health monitoring data, real-time operating parameters and environmental variables to achieve early warning of pressure vessel explosion accidents through high-precision risk prediction. Summary of the Invention

[0003] In order to solve the above problems, the present invention provides a pressure vessel explosion risk prediction method and prediction system.

[0004] The technical solution adopted in the present invention is:

[0005] A method for predicting the explosion risk of a pressure vessel comprises the following steps:

[0006] Step 1: Real-time collection of relevant data of the pressure vessel, including process parameters: pressure P and temperature T, structural parameters: wall thickness δ and surface crack length L, and environmental parameters: ambient humidity H and corrosive gas concentration C;

[0007] Step 2: Based on the relevant data collected in step 1, calculate the pressure fluctuation rate s and corrosion rate R of the pressure vessel c and residual strength S r ;

[0008] Step 3: Based on fracture mechanics theory and the relevant data collected in step 1, a critical failure pressure model of the pressure vessel is constructed, and the critical failure pressure is calculated according to the critical failure pressure model; a safety factor is calculated according to the critical failure pressure and the design pressure of the pressure volume; and a failure probability of the pressure vessel is obtained according to a safety factor-failure probability mapping table;

[0009] Step 4: Construct an XGBoost model based on the XGBoost algorithm, train the XGBoost model using historical data, input the pressure P, temperature T, wall thickness δ, surface crack length L collected in step 1, and the corrosion rate and residual strength calculated in step 2 into the trained XGBoost model, and obtain the predicted probability of the pressure vessel through XGBoost calculation;

[0010] Step 5: Based on the failure probability obtained in step 3 and the predicted probability obtained in step 4, a mixed risk probability of the pressure vessel is obtained after weight distribution;

[0011] Step 6: According to the mixed risk probability obtained in step 5, the pressure vessel is divided into different risk levels.

[0012] Furthermore, in step 2, the calculation formula of the pressure fluctuation rate s is as follows:

[0013]

[0014] Wherein, ΔP is the pressure change amplitude, in MPa; Δt is the pressure change time, in min.

[0015] Furthermore, in step 2, the corrosion rate R c The calculation formula is as follows:

[0016] R c =k×C×H 0.5 ;

[0017] Where k is the material corrosion coefficient of the pressure vessel under the average humidity throughout the year, C is the concentration of corrosive gas in ppm, and H is the ambient humidity.

[0018] Furthermore, in step 2, the residual strength S r The calculation formula is as follows:

[0019] S r =S0-α×L 2 ;

[0020] Where S0 is the initial design strength of the pressure vessel, α is the crack sensitivity coefficient, and L is the surface crack length, in mm.

[0021] Furthermore, in step 3, the calculation formula of the critical failure pressure model is as follows:

[0022]

[0023] Wherein, P is the critical failure pressure of the pressure vessel, in MPa; K ICis the fracture toughness of the pressure vessel material, in MPa·m; L is the surface crack length, in mm; δ is the wall thickness, in mm.

[0024] Furthermore, in step 3, the safety factor-failure probability mapping table is as follows:

[0025] Table 1

[0026] Safety factor n <![CDATA[Failure probability P crit > 45 10％ 35 30％ 25 50％ 15 70％ 5 90％

[0027] Furthermore, in step 4, the historical data are the pressure P, temperature T, wall thickness δ, surface crack length L, ambient humidity H and corrosive gas concentration C of the pressure vessel detected over a period of 3 years.

[0028] Furthermore, in step 5, the mixed risk probability P final The calculation formula is as follows:

[0029] P final =α×P crit +(1-α)×P data ;

[0030] Wherein, α is the weight coefficient, which is selected adaptively according to the stability of the working condition. In stable working condition, α = 0.7, and in transient working condition, α = 0.3. The working condition stability is judged according to the pressure fluctuation rate s calculated in step 2; P crit is the failure probability, P data is the predicted probability.

[0031] Furthermore, in step 6, the pressure vessel is classified into three risk levels: high, medium, and low according to the table below, and corresponding control measures are taken according to the risk levels.

[0032] Table 2

[0033] Mixed risk probability Risk Level Control measures P<0.3 Low Normal monitoring 0.3≤P<0.7 middle Strengthen inspections P≥0.7 high Emergency shutdown

[0034] A pressure vessel explosion risk prediction system comprises: a memory and a processor; wherein the memory stores a computer program, and when the computer program is executed by the processor, it can implement any one of the pressure vessel explosion risk prediction methods described above.

[0035] Beneficial effects of the present invention: The present invention breaks through the limitations of traditional reliance on manual detection and statistical models. By constructing a "physics-data" dual-driven dynamic prediction framework, the multi-dimensional data collected by sensors in real time is combined with fracture mechanics theory and the XGBoost algorithm to realize hybrid probability calculation and dynamic grading of pressure vessel explosion risks. It can not only capture dynamic risks and adapt to complex working conditions, but also significantly improve the accuracy and reliability of risk warnings, providing chemical companies with a more scientific early warning solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The figure is a flow chart of the method for predicting the explosion risk of a pressure vessel according to the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and preferred embodiments.

[0038] Example 1

[0039] See Figure 1 This application takes the (o)nitroaniline reactor of a chemical plant in Jiangxi as an example to provide a method for predicting the explosion risk of pressure vessels in a chemical enterprise, including the following steps:

[0040] Step 1: Use sensors to collect relevant data of key parts of the pressure vessel in real time, including process parameters (pressure P, temperature T), structural parameters (wall thickness δ, surface crack length L), and environmental parameters (ambient humidity H, corrosive gas concentration C).

[0041] In this embodiment, the method for collecting key data is as follows:

[0042] Dynamic monitoring of process parameters: Real-time data collection is achieved through a high-precision embedded sensor network. Explosion-proof pressure transmitters are deployed on the vessel body and inlet and outlet pipes to capture pressure fluctuations and transient shocks. Distributed fiber-optic temperature sensors are attached to the inner wall of the vessel to simultaneously monitor temperature gradients and localized overheating.

[0043] Non-destructive monitoring of structural parameters: Use fixed ultrasonic thickness gauges to be deployed in stress concentration areas (such as welds and openings), automatically scan and compensate for the effect of temperature on sound velocity at regular intervals; use an acoustic emission sensor network to capture elastic wave signals during crack propagation, and identify crack activity through wavelet transform.

[0044] Distributed sensing of environmental parameters: Explosion-proof capacitive humidity sensors are deployed on the outer wall and supporting structure of the container. Electrochemical gas sensors are placed in ventilation dead corners around the container, and the local concentration field is inverted through the gas diffusion model.

[0045] Step 2: Calculate the pressure fluctuation rate, corrosion rate, residual strength and other related derivative data of the reactor;

[0046] In the safety monitoring of chemical equipment, pressure fluctuation rate, corrosion rate and residual strength are key safety assessment parameters.

[0047] The pressure fluctuation rate represents the magnitude of pressure change per unit time and is used to reflect the dynamic stability of pressure within a device. Severe pressure fluctuations can accelerate material fatigue and lead to device failure. The magnitude of the pressure fluctuation rate can be used to determine whether the device is operating in a stable or transient state.

[0048] In this embodiment, the calculation formula of the pressure fluctuation rate s is:

[0049]

[0050] For example, the pressure of the reactor rises from 6.3 MPa to 7.52 MPa within 30 minutes. The pressure fluctuation rate is:

[0051]

[0052] In the chemical industry, if the pressure fluctuation rate s of a pressure vessel is less than 0.01 MPa / s, it is generally considered that the pressure vessel is in a stable operating condition. If it exceeds the range, it is considered that the pressure vessel is in a transient operating condition.

[0053] The corrosion rate i quantifies the rate of material loss under specific conditions and is a key indicator for predicting the remaining life of equipment. Its calculation formula is:

[0054] R c =k×C×H 0.5 ; where k is the material corrosion coefficient.

[0055] The para-(o-)nitroaniline reactor at this chemical plant is made of Q235B. Historical meteorological data indicates that the average annual humidity in the area is 75%. According to ISO 9223 (Classification of Atmospheric Corrosivity), the material corrosion coefficient for carbon steel in an industrial atmosphere with a humidity of 75% is 0.01. The monitored corrosive gas concentration is 50 ppm, and the ambient humidity is 75%. The corrosion rate is calculated as follows:

[0056] R c =0.01×50×75 0.5 =4.33mm.

[0057] Residual strength indicates the current load-bearing capacity of equipment with defects (such as cracks) and is used to assess structural safety. The calculation formula is as follows:

[0058] S r =S0-α×L 2 ;

[0059] Where S0 is the initial strength and α is the crack sensitivity coefficient.

[0060] The initial strength of the reactor is 235 MPa. The crack sensitivity coefficient depends on the carbon content of the steel. The lower the carbon content, the lower the crack sensitivity coefficient. The crack sensitivity coefficient of Q235B is 1.8. The monitored reactor body weld crack length is 3 mm. Its residual strength is calculated as follows:

[0061] S r =235MPa-1.8×(0.003m) 2 =234.9999838MPa.

[0062] Step 3: Based on fracture mechanics theory, construct the critical failure pressure model of the reactor. The calculation formula is as follows:

[0063]

[0064] Among them, K IC is the fracture toughness of the material;

[0065] The fracture toughness of carbon steel is 125 MPa·m (1 / 2) The average length of the weld cracks in the reactor body tested is 3 mm, and the average wall thickness of the reactor body is 36 mm. The critical failure pressure is calculated as follows:

[0066]

[0067] The design pressure of the reactor is 6.3 MPa, and its safety factor is calculated as follows:

[0068]

[0069] According to the safety factor-failure probability mapping table 1, the failure probability P of the reactor is obtained. crit is 55%.

[0070] Table 1

[0071] Safety factor n <![CDATA[Failure probability P crit > 45 10％ 35 30％ 25 50％ 15 70％ 5 90％

[0072] Safety factor n and failure probability P crit There is a linear relationship.

[0073] Step 4: Construct an XGBoost model based on the XGBoost algorithm. Use the historical 3-year monitoring data, the pressure vessel pressure P, temperature T, wall thickness δ, surface crack length L, ambient humidity H and corrosive gas concentration C to train the XGBoost model. Input {pressure P = 7.52MPa, temperature T = 502K, vessel wall thickness δ = 36mm, surface crack length L = 3mm, corrosion rate R C =4.33mm, residual strength S R=234.9999838MPa}, and the predicted probability P is obtained by XGBoost algorithm. data =53.61%.

[0074] Step 5: Based on the failure probability obtained in step 3 and the predicted probability obtained in step 4, the mixed risk probability of the pressure vessel is calculated according to the following formula;

[0075] To improve the accuracy and objectivity of predictions, a hybrid risk probability calculation method is proposed, combining data-driven (frequency) and physical models. This hybrid risk calculation method not only uses historical data but also incorporates physical models when calculating risk probability, thus compensating for the shortcomings of a single data source.

[0076] The formula for constructing the hybrid judgment matrix is:

[0077] P final =α×P crit +(1-α)×P data

[0078] Wherein, α is the weight coefficient, which is adaptively adjusted according to the stability of the working condition (α = 0.7 for stable working condition and α = 0.3 for transient working condition).

[0079] In this embodiment, according to the pressure fluctuation rate in step 2, it can be known that the reactor is in a stable working condition. Taking α = 0.7, step 3 calculates P crit =55%, step 4 calculates P data =53.61%, the mixed risk probability is calculated as follows:

[0080] P final =0.7×55%+(1-0.7)×53.61%=54.58%.

[0081] Step 6: Based on the mixed risk probability obtained in step 5 and the table below, classify the risk level of the reactor and formulate corresponding measures.

[0082] Table 2

[0083] Mixed risk probability Risk Level Control measures P<0.3 Low Normal monitoring 0.3≤P<0.7 middle Strengthen inspections P≥0.7 high Emergency shutdown

[0084] According to Table 2, the mixed risk probability of 54.58% in this embodiment is between 0.3 and 0.7. The risk level of the reactor is medium, and inspections should be strengthened.

[0085] Example 2

[0086] The present application provides a pressure vessel explosion risk warning system, comprising: a memory and a processor; wherein the memory stores a computer program, and when the program is executed by the processor, it can implement the pressure vessel explosion risk warning method as described in Example 1.

[0087] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also within the scope of protection of the present invention.

Claims

1. A method for predicting the risk of explosion of a pressure vessel, characterized in that: The following steps are involved: Step 1: Real-time collection of relevant data of the pressure vessel, including process parameters: pressure P and temperature T, structural parameters: wall thickness δ and surface crack length L, and environmental parameters: ambient humidity H and corrosive gas concentration C; Step 2: Based on the relevant data collected in step 1, calculate the pressure fluctuation rate s and corrosion rate R of the pressure vessel c and residual strength S r ; Step 3: Based on fracture mechanics theory and the relevant data collected in step 1, a critical failure pressure model of the pressure vessel is constructed, and the critical failure pressure is calculated according to the critical failure pressure model; a safety factor is calculated according to the critical failure pressure and the design pressure of the pressure volume; and a failure probability of the pressure vessel is obtained according to a safety factor-failure probability mapping table; Step 4: Construct an XGBoost model based on the XGBoost algorithm, train the XGBoost model using historical data, input the pressure P, temperature T, wall thickness δ, surface crack length L collected in step 1, and the corrosion rate and residual strength calculated in step 2 into the trained XGBoost model, and obtain the predicted probability of the pressure vessel through XGBoost calculation; Step 5: Based on the failure probability obtained in step 3 and the predicted probability obtained in step 4, a mixed risk probability of the pressure vessel is obtained after weight distribution; Step 6: According to the mixed risk probability obtained in step 5, the pressure vessel is divided into different risk levels.

2. The method for predicting the risk of explosion of a pressure vessel according to claim 1, characterized in that: In step 2, the calculation formula of the pressure fluctuation rate s is as follows: Wherein, ΔP is the pressure change amplitude, in MPa; Δt is the pressure change time, in min.

3. The method for predicting the explosion risk of a pressure vessel according to claim 1, wherein: In step 2, the corrosion rate R c The calculation formula is as follows: R c =k×C×H 0.5 ; Where k is the material corrosion coefficient of the pressure vessel under the average humidity throughout the year, C is the concentration of corrosive gas in ppm, and H is the ambient humidity.

4. The method for predicting the explosion risk of a pressure vessel according to claim 1, wherein: In step 2, the residual strength S r The calculation formula is as follows: S r =S0-α×L 2 ; Where S0 is the initial design strength of the pressure vessel, α is the crack sensitivity coefficient, and L is the surface crack length, in mm.

5. The method for predicting the risk of explosion of a pressure vessel according to claim 1, characterized in that: In step 3, the calculation formula of the critical failure pressure model is as follows: Wherein, P is the critical failure pressure of the pressure vessel, in MPa; K IC is the fracture toughness of the pressure vessel material, in MPa·m; L is the surface crack length, in mm; δ is the wall thickness, in mm.

6. The method for predicting the explosion risk of a pressure vessel according to claim 1, wherein: In step 3, the safety factor-failure probability mapping table is as follows: Table 1 7. The method for predicting the risk of explosion of a pressure vessel according to claim 1, characterized in that: In step 4, the historical data are the pressure P, temperature T, wall thickness δ, surface crack length L, ambient humidity H, and corrosive gas concentration C of the pressure vessel detected over a period of three years.

8. The method for predicting the explosion risk of a pressure vessel according to claim 1, wherein: In step 5, the mixed risk probability P final The calculation formula is as follows: P final =α×P crit +(1-a)×P data ; Wherein, α is the weight coefficient, which is selected adaptively according to the stability of the working condition. In stable working condition, α = 0.7, and in transient working condition, α = 0.

3. The working condition stability is judged according to the pressure fluctuation rate s calculated in step 2; P crit is the failure probability, P data is the predicted probability.

9. The method for predicting the explosion risk of a pressure vessel according to claim 1, wherein: In step 6, the pressure vessel is classified into three risk levels: high, medium, and low according to the table below, and corresponding control measures are taken according to the risk levels; Table 2 10. A pressure vessel explosion risk prediction system, characterized in that: include: A memory and a processor; wherein the memory stores a computer program, and when the computer program is executed by the processor, it can implement the pressure vessel explosion risk prediction method according to any one of claims 1 to 9.

Images