Automatic annealing production line for pressure vessels

Abstract

This invention discloses an automated annealing production line for pressure vessels, achieving a comprehensive upgrade of the pressure vessel annealing process through multi-dimensional technology integration. In terms of temperature control, it employs a combination of real-time infrared thermal imaging monitoring and electromagnetic dynamic compensation to precisely adjust the heating power of each zone, completely solving the problem of excessive temperature differences caused by uneven heat conduction in traditional processes. For the cooling stage, an innovative zoned frequency conversion control system based on wall thickness detection has been developed. This system automatically matches the optimal cooling rate according to the thickness of different parts of the vessel, effectively suppressing the risk of deformation and cracking. The system's built-in self-learning algorithm continuously optimizes the process parameter library, significantly reducing the trial-and-error costs of new product development. Simultaneously, the fully automated design enables intelligent operation from material loading and unloading, temperature control to quality prediction, minimizing manual intervention.

Description

Technical Field

[0001] This invention relates to the field of pressure vessel annealing technology, specifically to an automatic pressure vessel heating annealing production line. Background Technology

[0002] Currently, pressure vessel annealing commonly employs gas-fired radiant heating or resistance furnace heating, which suffers from core bottlenecks such as low heat transfer efficiency and uneven temperature distribution. Due to differences in the thermal inertia of metallic materials, the heat accumulation effect is significant in thick-walled areas, weld locations, and areas of structural abrupt changes, resulting in actual temperature differences of up to 50-80℃ during the heating / holding stages (far exceeding the ±25℃ requirement of the ASME standard). This non-uniformity leads to two serious consequences: firstly, grain coarsening occurs in locally overheated areas (such as the middle of the cylinder), reducing the material's yield strength by 15%-20%; secondly, insufficient stress release occurs in low-temperature areas (such as the transition zone of the end cap), with residual stress exceeding 200MPa, creating a potential risk of stress corrosion cracking in the pressure vessel. Even more serious is the lack of a real-time temperature compensation mechanism in existing processes; operators can only manually adjust the gas valve through the furnace door observation port, resulting in a response lag of over 15 minutes, which fails to effectively suppress temperature fluctuations.

[0003] Mainstream annealing production lines are still in the transition phase from mechanization to semi-automation, with key processes relying on manual operation: container transfer requires a bridge crane and manual positioning, with an average loading and unloading time exceeding 40 minutes; the cooling process uses overall forced draft or natural air cooling, making it impossible to implement gradient control based on wall thickness differences, resulting in excessively rapid cooling rates in thin-walled areas (>50℃ / min), producing martensitic phase transformation hard spots, while thick-walled areas are insufficiently cooled (<10℃ / min), extending the production cycle. According to NACE International statistics, the resulting container deformation and rework rate is as high as 12%-18%. At the control level, the PLC system only implements basic temperature control loops, and process parameter settings rely entirely on worker experience. For example, the heating power is fixed at 80%-100% of the rated value, unable to dynamically match the container's thermal state; the cooling fan only supports start-stop control, lacking linear speed adjustment functionality. This rigid control mode leads to significant energy waste, with single-piece annealing gas consumption reaching 0.8-1.2 tons of standard coal.

[0004] Current technologies lack end-to-end data closed-loop capabilities: temperature data is collected via limited thermocouples (typically <10 measurement points), making it impossible to construct a complete thermal field model; a quantitative correlation has not been established between cooling parameters and final hardness / metallographic structure. A more prominent problem is that process parameter adjustments rely on post-process destructive testing (such as cutting samples for hardness testing), with feedback cycles as long as 24-48 hours, leading to frequent batch quality incidents. In 2019, a pressure vessel factory scrapped an entire batch of products due to excessive head hardness, resulting in direct losses exceeding 6 million yuan. Furthermore, experience-based parameter databases are stored in paper documents or scattered spreadsheets, failing to achieve intelligent matching based on material properties and wall thickness specifications. When processing new vessel models, 3-5 rounds of trial-and-error annealing are required, with an average debugging cost of 200,000 yuan per instance.

[0005] Although some companies have attempted to introduce automated equipment (such as robotic arms for loading and unloading) or simulation software (such as ANSYS thermal analysis), the problem of technological silos exists: robotic arms are not linked to thermal field monitoring and only replace manual handling; simulation results are not fed back to the control system in real time, becoming offline verification tools. Recent research on intelligent annealing focuses on single technical points and has not solved the system engineering challenge of thermo-mechanical-control coordination. Summary of the Invention

[0006] To address or partially address the problems existing in related technologies, this invention application provides an automatic annealing production line for pressure vessels, comprising: A preheating device with zoned electromagnetic coils; Gas-fired main annealing furnace; Variable frequency fan cooling chamber; Central controller, execute: a) Real-time temperature field analysis algorithm b) Dynamic Allocation Algorithm for Hybrid Heating Power c) Process parameter self-learning and updating module.

[0007] Furthermore, the real-time temperature field analysis algorithm acquires the container surface temperature matrix using an infrared thermal imager, identifies the coordinates of the low-temperature zone and the temperature difference value ΔT, and outputs a compensation priority sequence.

[0008] Furthermore, the dynamic allocation algorithm for hybrid heating power is adjusted according to the following rules: When ΔT≤20℃, the power ratio of gas heating is ≥80%. When ΔT>20℃, the electromagnetic compensation power is increased according to the formula: \( P_{em} = K \cdot \Delta T \cdot \log_{10}(A_{low} / A_{total}) \) Where K is the material coefficient, A lowA represents the area of ​​the low-temperature region. total For the total area Furthermore, the algorithm applies reinforcement compensation to the weld area: Weld coordinates are located using image recognition; When the temperature of the weld zone ΔT ≥ 15℃, an additional 20% electromagnetic power is required; After compensation, the temperature difference control accuracy reaches ±8℃.

[0009] Furthermore, the cooling chamber control strategy includes: Cooling zones were defined based on wall thickness measurement data; Set the fan base speed according to the thickness gradient d: \(RPM_{base} = 1500 / (1+0.1d) \) The rotation speed is dynamically adjusted by ±15% based on the cooling rate; Furthermore, the process parameter self-learning update module executes: Store the temperature profile, power distribution, cooling parameters, and final hardness for each annealing process; When the hardness deviation ΔHV ≥ 8: a) Automatic adjustment of the K coefficient: \( K_{new} = K_{old} \times (1 + 0.05 \cdot sign(\Delta HV)) \) b) Update the priority sequence weights.

[0010] Furthermore, the module establishes a material-thickness-parameter mapping table for processing new workpieces: Retrieve historical records for material deviation ≤5% and thickness deviation ≤10%; Automatically load the optimal initial parameters.

[0011] Furthermore, it includes a quality prediction and alarm unit: Real-time calculation of temperature uniformity index\( U_t = \frac{\max(T)-\min(T)}{\text{avg}(T)}\) WhenU t An audible and visual alarm is triggered when the value is >0.25. Predict the final hardness value and display it on the HMI interface.

[0012] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit the scope of this application.

[0013] Beneficial technical effects of the present invention: This application integrates multiple technologies to achieve a comprehensive upgrade of the pressure vessel annealing process. In terms of temperature control, it employs a combination of real-time infrared thermal imaging monitoring and electromagnetic dynamic compensation to precisely adjust the heating power of each zone, completely resolving the problem of excessive temperature differences caused by uneven heat conduction in traditional processes. For the cooling stage, an innovative zoned frequency conversion control system based on wall thickness detection has been developed. This system automatically matches the optimal cooling rate according to the thickness of different parts of the vessel, effectively suppressing the risk of deformation and cracking. The system's built-in self-learning algorithm continuously optimizes the process parameter library, significantly reducing the trial-and-error costs of new product development. In terms of energy efficiency management, the synergistic cooperation between gas-fired main heating and precise electromagnetic compensation, along with a waste heat recovery system, significantly reduces energy consumption. Simultaneously, the fully automated design enables intelligent operation from material loading and unloading, temperature control to quality prediction, minimizing manual intervention. These technological innovations work together to construct a highly efficient, precise, energy-saving, and intelligent pressure vessel annealing production system, significantly reducing production costs while improving product quality. Attached Figure Description

[0014] Figure 1 This is a control system architecture diagram of one embodiment of the present invention. Detailed Implementation

[0015] The alternative embodiments of this application will now be described in more detail with reference to the accompanying drawings. While alternative embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to make this application more thorough and complete, and to fully convey the scope of this application to those skilled in the art.

[0016] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0017] The automatic annealing production line for pressure vessels according to this invention will be described in detail below with reference to the accompanying drawings: include: A preheating device with zoned electromagnetic coils; Gas-fired main annealing furnace; Variable frequency fan cooling chamber; Central controller, execute: a) Real-time temperature field analysis algorithm b) Dynamic Allocation Algorithm for Hybrid Heating Power c) Process parameter self-learning and updating module.

[0018] In one embodiment of this invention application, the real-time temperature field analysis algorithm acquires the surface temperature matrix of the container using an infrared thermal imager, identifies the coordinates of the low-temperature zone and the temperature difference value ΔT, and outputs a compensation priority sequence.

[0019] Infrared thermal imagers convert temperature distribution into matrix data, identify continuous low-temperature areas through edge detection algorithms, and generate compensation priority queues based on temperature difference ΔT and area (e.g., areas with ΔT>30℃ are processed first).

[0020] In one embodiment of this invention, the dynamic allocation algorithm for hybrid heating power is adjusted according to the following rules: When ΔT≤20℃, the power ratio of gas heating is ≥80%. When ΔT>20℃, the electromagnetic compensation power is increased according to the formula: \( P_{em} = K \cdot \Delta T \cdot \log_{10}(A_{low} / A_{total}) \) Where K is the material coefficient, A low A represents the area of ​​the low-temperature region. total For the total area When the temperature difference is small (ΔT≤20℃), gas heating is the primary method (high energy efficiency). When there is a large temperature difference, the electromagnetic compensation power is increased according to the formula \( P_{em} = K \cdot \Delta T \cdot \log_{10}(A_{low} / A_{total}) \), where the logarithmic function suppresses ineffective compensation (high power consumption is not triggered in small low-temperature areas). In one embodiment of this invention, the algorithm performs reinforcement compensation on the weld area: Weld coordinates are located using image recognition; When the temperature of the weld zone ΔT ≥ 15℃, an additional 20% electromagnetic power is required; After compensation, the temperature difference control accuracy reaches ±8℃.

[0021] Identifying weld seam image features (straight-line high-temperature bands) using the OpenCV library When the weld temperature ΔT ≥ 15℃, add 20% power to the basic compensation amount. In one embodiment of this invention application, the cooling chamber control strategy includes: Cooling zones were defined based on wall thickness measurement data; Set the fan base speed according to the thickness gradient d: \(RPM_{base} = 1500 / (1+0.1d) \) The rotation speed is dynamically adjusted by ±15% based on the cooling rate; Substitute the wall thickness d into the empirical formula \( RPM_{base} = 1500 / (1+0.1d) \) to calculate the base rotation speed (for thick walls, reduce speed to prevent cracking). Real-time monitoring of cooling rate; automatic adjustment of ±15% if the rate exceeds the limit (if the rate is <2℃ / min, the speed will be increased). In one embodiment of this invention application, the process parameter self-learning update module performs the following: Store the temperature profile, power distribution, cooling parameters, and final hardness for each annealing process; When the hardness deviation ΔHV ≥ 8: a) Automatic adjustment of the K coefficient: \( K_{new} = K_{old} \times (1 + 0.05 \cdot sign(\Delta HV)) \) b) Update the priority sequence weights.

[0022] When the hardness deviation ΔHV ≥ 8, a correction factor is applied (the sign function takes the direction of the hardness deviation). Update the mapping table weights (e.g., if a material parameter is called 10 times and the average ΔHV value is <5, increase the weight by 20%). In one embodiment of this invention, the module establishes a material-thickness-parameter mapping table. When processing a new workpiece: Retrieve historical records for material deviation ≤5% and thickness deviation ≤10%; Automatically load the optimal initial parameters.

[0023] Establish a 3D parameter mapping table (material × thickness × optimal parameters). Similarity retrieval uses the Euclidean distance algorithm: \( \sqrt{(material deviation)^2 + (thickness deviation)^2} \) In one embodiment of this invention application, a quality prediction and alarm unit is included: Real-time calculation of temperature uniformity index\( U_t = \frac{\max(T)-\min(T)}{\text{avg}(T)}\) WhenU t An audible and visual alarm is triggered when the value is >0.25. Predict the final hardness value and display it on the HMI interface.

[0024] Temperature uniformity index \( U_t = \frac{\max(T)-\min(T)}{\text{avg}(T)} \) is calculated in real time (an alarm is triggered when the index is >0.25). Hardness prediction is based on a regression model: \( HV_{pred} = a \cdot U_t + b \cdot \text{cooling rate} + c \) For clarity, the following examples will be used to provide a detailed description.

[0025] Example 1: System Hardware Configuration 1. Basic Structure Preheating device: It adopts 3 sets of independently controlled 50kW medium frequency electromagnetic coils (model EF-50, frequency 1-10kHz), and the coil spacing is adjustable (300-800mm). Main annealing furnace: Gas-fired radiant tube (maximum power 120kW) equipped with 6 sets of embedded 20kW electromagnetic compensation coils (arranged in the corresponding areas of the furnace door / weld seam). Cooling chamber: 8 x 5.5kW variable frequency centrifugal fans (wind speed range 2-15m / s), arranged axially in two layers. Detection unit: Infrared thermal imager (FLIR A615, temperature measurement accuracy ±2℃) Handheld laser thickness gauge (accuracy ±0.1mm, scanning 5 points along the container axis) Example 2: Annealing Control Process Step 1: Initial parameter settings Input container material (Q345R), design wall thickness (12mm), and target temperature (650±15℃). The self-learning module retrieves historical databases and loads initial parameters for similar processes. If material matching accuracy > 90% and thickness deviation < 10%: Call {K=0.32, RPM_base=1200, T_target=650} else: Using default parameters {K=0.3, RPM_base=1000} Step 2: Dynamic Power Allocation The infrared thermal imager acquires a temperature matrix every 30 seconds (resolution 640×480). Identify low-temperature areas (continuous pixel areas with temperatures below the average value). Calculate electromagnetic compensation power: When the temperature ΔT = 25℃ (weld zone), A_low / A_total = 0.18 (material coefficient K = 0.32). P_em = 0.32 × 25 × log10(0.18) = 0.32×25×(-0.744) ≈ 5.95kW Weld reinforcement compensation: Additional 20% → Total P_em = 7.14kW The gas valve is controlled to maintain a base power of 96kW (93% of the total power). Step 3: Gradient Cooling Control Laser thickness gauge obtains wall thickness data: cylinder 12mm, end cap 15mm Calculate the fan speed using the formula: Cylinder region: RPM_base = 1500 / (1+0.1×12) = 682 Header area: RPM_base = 1500 / (1+0.1×15) = 600 When the head cooling rate is less than 3℃ / min, the speed is automatically increased to 690 RPM (+15%). Step 4: Self-learning update The weld hardness was tested after annealing: measured HV185 (target value HV190). Triggering update mechanism: ΔHV = |185-190| = 5<8 → No coefficient update If ΔHV≥8 (e.g., measured HV180): K_new = K_old × (1 + 0.05×sign(-5)) = 0.32×0.95 = 0.304 The new parameters are stored in the process library and marked as

Q345R_12mm_v2

[0026] Technical effect verification data The embodiments of this invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. An automatic annealing production line for pressure vessels, characterized in that, include: A preheating device with zoned electromagnetic coils; Gas-fired main annealing furnace; Variable frequency fan cooling chamber; Central controller, execute: a) Real-time temperature field analysis algorithm, b) Dynamic allocation algorithm for hybrid heating power. c) Process parameter self-learning and updating module.

2. The production line as described in claim 1, characterized in that: The real-time temperature field analysis algorithm acquires the container surface temperature matrix using an infrared thermal imager, identifies the coordinates of the low-temperature zone and the temperature difference value ΔT, and outputs a compensation priority sequence.

3. The production line as described in claim 2, characterized in that: The dynamic allocation algorithm for hybrid heating power is adjusted according to the following rules: When ΔT≤20℃, the power ratio of gas heating is ≥80%. When ΔT>20℃, the electromagnetic compensation power is increased according to the formula: \( P_{em} = K \cdot \Delta T \cdot \log_{10}(A_{low} / A_{total}) \) Where K is the material coefficient, A low A represents the area of ​​the low-temperature region. total For the total area The production line as described in claim 3 is characterized in that: The algorithm applies reinforcement compensation to the weld area: Weld coordinates are located using image recognition; When the temperature of the weld zone ΔT ≥ 15℃, an additional 20% electromagnetic power is required; After compensation, the temperature difference control accuracy reaches ±8℃.

4. The production line as described in claim 1, characterized in that: Cooling chamber control strategies include: Cooling zones were defined based on wall thickness measurement data; Set the fan base speed according to the thickness gradient d: \(RPM_{base} = 1500 / (1+0.1d) \) The rotation speed is dynamically adjusted by ±15% based on the cooling rate; The production line as described in claim 1, characterized in that: The process parameter self-learning update module is executed as follows: Store the temperature profile, power distribution, cooling parameters, and final hardness for each annealing process; When the hardness deviation ΔHV ≥ 8: a) Automatically adjust the K coefficient: \( K_{new} = K_{old} \times (1 + 0.05 \cdot sign(\DeltaHV)) \) b) Update the priority sequence weights.

5. The production line as described in claim 6, characterized in that: The module establishes a material-thickness-parameter mapping table. When machining a new workpiece: Retrieve historical records for material deviation ≤5% and thickness deviation ≤10%; Automatically load the optimal initial parameters.

6. The production line as described in claim 1, characterized in that: Includes a quality prediction and alarm unit: Real-time calculation of temperature uniformity index \( U_t = \frac{\max(T)-\min(T)}{\text{avg}(T)} \) When U t An audible and visual alarm is triggered when the value is >0.

25. Predict the final hardness value and display it on the HMI interface.

Images