Intelligent control method and system for local heating temperature field of tank body weld

Abstract

The application discloses a kind of tank weld local heat treatment temperature field intelligent regulation and control method and system, it is related to data processing technical field, method includes: ring slit or longitudinal slit is divided into weld regulation and control section, establishes temperature mapping table and generates development temperature point, identifies the width of double-sided heat affected zone, with the previous adjacent section and dynamic boundary data comparison generates boundary deviation data and section-level revisit data, according to this generates heat treatment state code, filters out the candidate process parameters of opposite heat input direction, obtains and executes target process parameter, again according to execution residual data correction subsequent control action and dynamic boundary data.The application solves the problem that the direction of parameter adjustment is unstable due to the mixing of heat accumulation between sections and double-sided boundary deviation in continuous welding, improves the consistency and reviewability of local heat treatment regulation and control of weld.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for intelligent control of the local temperature field of tank weld homogenization. Background Technology

[0002] Tank welds are widely found in pressure vessels, storage tanks, reaction vessels, transport tanks, and large cylindrical structures. These welds typically include circumferential welds extending circumferentially along the tank body and longitudinal welds extending axially. Compared to short welds on flat surfaces, tank welds are located on curved surfaces, where the welding heat source moves continuously across the curved surface, resulting in continuous heat transfer between the welded area, the area being welded, and the area awaiting welding. Furthermore, the weld centerline is affected by plate thickness, bevel condition, fixture contact, tank support position, and temperature measurement projection errors, causing the heat-affected zone boundary to not always unfold symmetrically along both sides of the weld centerline.

[0003] Current welding temperature field control typically employs methods such as temperature point boundary judgment, regional average temperature judgment, temperature gradient judgment, or overall temperature field prediction. While these methods can respond to localized overheating or abnormal temperature gradients, their primary control object remains the temperature value or gradient, making it difficult to directly reflect the boundary location of the heat-affected zone (HAZ), which is more crucial for tank weld quality control. For example, in the case of a tank circumferential weld, as the welding heat source operates continuously along the circumference, residual heat from previous weld segments is transferred along the tank's curved surface to the current weld segment. If only the temperature at the current sampling point is observed, the boundary expansion caused by previous heat accumulation might be mistaken for a current instantaneous disturbance, leading to inappropriate adjustments. Similarly, in the case of a tank longitudinal weld, if the first side of the weld centerline is close to the fixture contact heat dissipation area, while the second side is in a free heat dissipation area, the HAZ boundaries on both sides of the weld may exhibit unilateral expansion or insufficiency. If only overall heat input adjustment is used, one boundary might be corrected while the other deviates from the target range.

[0004] Furthermore, existing closed-loop feedback mechanisms often rely on whether the temperature drops or returns to within acceptable limits after execution, lacking a deterministic comparison between pre-execution boundary deviations, theoretical responses to target process parameters, and actual post-execution boundary responses. After a change in welding heat input, the heat-affected zone boundary does not immediately return to its original state but is influenced by the combined effects of power supply response, motion response, heat conduction hysteresis, and the thermal history of the weld segment. Without a comparison of pre- and post-execution boundary responses, it is easy to mistakenly attribute the execution link's failure to take effect to insufficient parameter adjustment, or to conclude that the dynamic boundary data is mismatched due to incorrect welding process parameters. Summary of the Invention

[0005] In response to the technical problems described in the background art, the present invention provides a method and system for intelligent control of local heat equalization temperature field of tank welds, which is used to solve the problem that in continuous welding of circumferential or longitudinal welds of tanks, the boundary deviation of the double-sided heat-affected zone of the current weld control section is mixed with the heat accumulation continuity relationship of the previous adjacent weld control section, resulting in unstable adjustment direction of process parameters.

[0006] A method for intelligent control of the local heat homogenization temperature field of a tank weld includes: dividing the circumferential weld of the tank according to the circumferential arc length or the longitudinal weld of the tank according to the axial length into weld control segments, and writing the temperature measurement point number, unfolding position, and weld centerline side into a temperature measurement mapping table; converting the temperature measurement value of the current sampling cycle into unfolding temperature points with segment number, side, unfolding position, and reliability mark according to the temperature measurement mapping table; within the same weld control segment, reading the first side temperature point and the second side temperature point along the direction away from the weld centerline to obtain the width of the double-sided heat-affected zone; and calculating the width of the double-sided heat-affected zone of the current weld control segment. The boundary deviation data and segment-level backtracking data are obtained by comparing the width on the same side and the corresponding dynamic boundary data of the adjacent weld control segment in the preceding sequence. A heat-aware status code is generated from the boundary deviation data and the segment-level backtracking data. Candidate process parameters with opposite directions are screened out according to the heat input direction defined by the heat-aware status code. The target process parameters are obtained by combining the boundary deviation data of the subsequent adjacent weld control segment. After executing the target process parameters, the width of the heat-affected zone on both sides after execution is compared with the theoretical response data to obtain the execution residual data. The subsequent control actions and dynamic boundary data are corrected by the execution residual data.

[0007] The beneficial effects of this invention are reflected in: In the intelligent control method for local heat homogenization temperature field of tank welds, the local heat homogenization control of tank welds is adjusted from temperature point control, temperature gradient control, or overall heat input control to a process based on weld control segments, temperature mapping tables, unfolded temperature points, width of the double-sided heat-affected zone, segment-level back-visit data, heat homogenization status codes, and execution residual data. Through weld control segments, circumferential or longitudinal welds are divided into processing objects that can be compared individually, revisited adjacently, and recorded independently. Through temperature mapping tables, temperature values ​​on the tank surface are converted to a unified local unfolded position of the weld. Through the width of the double-sided heat-affected zone, the boundary expansion, insufficient boundary, or double-sided imbalance of the first and second sides of the weld centerline are recorded respectively. Through segment-level back-visit data, the local offset of the current weld control segment is distinguished from the heat accumulation formed by the continuation of the previous adjacent weld control segment. Through heat input direction filtering, candidate process parameters opposite to the boundary convergence direction are excluded. Through post-execution boundary response comparison, convergence in the same direction, insufficient response, back-readback failure, and reverse divergence are recorded. The entire scheme enables the process of local heat equalization control of tank welds to be recorded segment by segment, compared side by side, and corrected based on the execution results. Attached Figure Description

[0008] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0009] Figure 1 This is a schematic diagram illustrating the steps of the intelligent control method for local heat equalization temperature field of tank weld seams according to the present invention. Figure 2 This is a schematic diagram of some steps in S1 of the intelligent control method for local heat equalization temperature field of tank weld in the present invention. Figure 3 This is a schematic diagram of the data flow timing in S2 of the intelligent control method for local heat equalization temperature field of tank weld in this invention; Figure 4 This is a schematic diagram of part of step S2 in the intelligent control method for local heat equalization temperature field of tank weld of the present invention; Figure 5 This is a schematic diagram of the data flow timing in S3 of the intelligent control method for local heat equalization temperature field of tank weld in this invention; Figure 6 This is a schematic diagram of some steps in S3 of the intelligent control method for local heat equalization temperature field of tank weld in this invention. Figure 7 This is a schematic diagram of the data flow timing in S4 of the intelligent control method for local heat equalization temperature field of tank weld in this invention; Figure 8 This is a schematic diagram of part of step S4 in the intelligent control method for local heat equalization temperature field of tank weld of the present invention; Figure 9 This is a schematic diagram of the module structure and data flow of the intelligent control system for local heat equalization temperature field of tank weld seam according to the present invention. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0011] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0012] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0013] This invention provides a method for intelligent control of the local heat equalization temperature field of tank welds, such as... Figure 1 As shown, in one embodiment, the method includes: S1. Divide the tank circumferential seam into weld control sections according to the circumferential arc length or the tank longitudinal seam according to the axial length. Write the temperature measurement point number, unfolding position and weld centerline side into the temperature measurement mapping table. Then, according to the temperature measurement mapping table, convert the temperature measurement value of the current sampling cycle into unfolding temperature points with section number, side, unfolding position and confidence mark. S2. Within the same weld control section, read the temperature points on the first and second sides along the direction away from the weld centerline to obtain the width of the heat-affected zone on both sides. Compare the width of the heat-affected zone on both sides of the current weld control section with the width on the same side and the corresponding dynamic boundary data of the adjacent weld control section to obtain boundary deviation data and section-level back-visit data. Generate a heat-equalizing status code from the boundary deviation data and section-level back-visit data. S3. Candidate process parameters with opposite directions are screened out according to the heat input direction defined by the heat homogenization status code, and the target process parameters are obtained by combining the boundary deviation data of the subsequent adjacent weld control sections. S4. After executing the target process parameters, compare the width of the heat-affected zone on both sides after execution with the theoretical response data to obtain the execution residual data, and use the execution residual data to correct subsequent control actions and dynamic boundary data.

[0014] In this embodiment, the weld seam control segment refers to the segment-level processing object obtained by dividing the circumferential or longitudinal seam of the tank along the welding direction. The corresponding general technical object is a continuous spatial segment on the weld seam path. The input data includes tank geometric data, weld seam path data, and welding direction. The generation method is to divide and record the segment number and the start and end range of the segment according to the circumferential arc length or axial length. The output results include the segment number, segment start point, segment end point, segment center position, previous adjacent segment number, and subsequent adjacent segment number. The unit or type includes length unit and segment number identifier. The usage stages include temperature measurement mapping, boundary identification, segment-level comparison, and execution residual recording. The technical function is to enable the temperature field control of the tank weld seam to have a fixed segment-level index. The unfolded position refers to the position of the temperature measuring point in the local unfolded coordinates of the weld. Its corresponding general technical object is the spatial coordinates of the temperature measuring point. The input data includes the temperature measuring position, the weld centerline position, the weld type, and the weld control section data. The generation method is to project the temperature measuring position of the curved surface of the tank onto the local coordinates of the weld and record the position along the weld direction and the position perpendicular to the weld direction. The output result is the circumferential arc length position or axial position and the position perpendicular to the weld direction, in units of length. The usage stages are the generation of unfolded temperature points and the identification of heat-affected zone boundaries. The technical function is to enable the temperature measuring points on the curved surface to be compared under the same local coordinates of the weld. A reliable temperature point marker is an identifier used to distinguish whether a deployed temperature point can participate in boundary identification. The corresponding general technical object is the temperature sampling quality identifier. The input data includes the measured temperature value, the measured temperature range, the sampling continuity record, the temperature mapping table, and the deployment position. The generation method is to determine whether the measured temperature value is within the range, whether the sampling before and after is continuous, and whether the deployment position is located in the effective area. The output results include the first reliable temperature point, the second reliable temperature point, and the invalid temperature point. The type is an enumeration identifier. The usage stage is the identification of the width of the bilateral heat-affected zone and the generation of dynamic boundary data. The technical function is to reduce the impact of missing sampling, out-of-boundary sampling, and mapping failure on boundary judgment.

[0015] In S1, the curved weld seam of the tank body is first converted into a data object that can be segmented and recorded, mapped to temperature measurement points, and formed into locally unfolded coordinates. The tank's geometric data includes the tank radius, tank plate thickness, tank axial length, and tank support position. The weld path data includes the weld type, weld centerline position, weld start position, weld end position, and welding direction. Taking the circumferential weld of a carbon steel storage tank as an example, the tank radius is 1800mm, the tank plate thickness is 12mm, the tank axial length is 6000mm, and the circumferential length of the circumferential weld is 1.131 × 10⁻⁶ mm. 4The circumferential length of the circumferential seam and the circumference corresponding to the tank radius are within the engineering rounding range and there is no geometric contradiction. The welding direction is clockwise. The first 47 weld control segments are divided according to the commonly used segment length of 235mm. The 48th weld control segment is used as a reserve segment with a length of approximately 265mm. The circumferential starting point of segment number 18 is 3995mm, the circumferential ending point is 4230mm, and the center position of the segment is 4112.5mm. After the temperature measurement point number, temperature measurement position, temperature measurement direction, and temperature measurement coverage area are written into the temperature measurement mapping table, the temperature measurement point number 035 is fixed to correspond to segment number 18, the first side, the circumferential arc length position of 4100mm, and the position perpendicular to the weld direction of 6mm. When the temperature measurement value of temperature measurement point number 035 is 812℃ within the sampling period, this temperature measurement value is no longer used as an isolated temperature value, but is converted into an unfolded temperature point with segment number, side, unfolded position, and reliable mark. This process ensures that the surface temperature measurement location, weld segment location, and subsequent boundary identification location all share the same index source.

[0016] In S2, within the same weld control section, the unfolded temperature points are first divided into first-side temperature points, second-side temperature points, and temperature points near the centerline, based on the weld centerline. Then, the temperature distribution changes are read along the direction away from the weld centerline to obtain the widths of the first-side and second-side heat-affected zones. Boundary points are not directly generated from dynamic boundary data but are determined by the distribution relationship of the current temperature points at their local unfolded positions; dynamic boundary data is only used for deviation comparison after the current heat-affected zone width is determined. Taking section number 18 as an example, the temperature values ​​measured at positions 6mm, 8mm, 10.4mm, and 12mm perpendicular to the weld on the first side are 812℃, 780℃, 742℃, and 690℃, respectively. The temperature decreases as it moves away from the weld centerline, consistent with the temperature distribution pattern near the welding heat source. Historical records of heat-affected zone edge temperatures in qualified welds range from 735℃ to 746℃. This range is close to the phase transformation temperature commonly used at the edge of the heat-affected zone in carbon steel welds, making it usable as a boundary temperature record for similar materials, plate thicknesses, and processes. Since the temperature measurement of 742℃ at the 10.4mm position is within the edge temperature record, and the temperature measurement of 690℃ at the 12mm position is below the lower limit of the edge temperature record, the boundary point of the first side heat-affected zone is recorded at the 10.4mm position, and the width of the first side heat-affected zone is 10.4mm. The second side is processed in the same way, resulting in a second side heat-affected zone width of 8.9mm. The effective stability window gives the first side dynamic boundary data of 8.8mm to 9.1mm and the second side dynamic boundary data of 8.7mm to 9.2mm. Therefore, the deviation direction of the first side is outward expansion, and the first side offset distance is 1.3mm. The deviation direction of the second side is normal. By reading the first side heat-affected zone width of 10.0mm and the second side heat-affected zone width of 8.8mm of the adjacent segment number 17, the segment-level back-visit data of the heat accumulation continuation on the first side can be obtained.

[0017] In S3, after generating the soaking status code from boundary deviation data and segment-level backtracking data, the process parameters are not changed directly. Instead, candidate process parameters with opposite directions are first filtered out according to the heat input direction defined by the soaking status code. The candidate process parameters are derived from the current welding process parameters, the equipment's minimum adjustment unit, historical valid execution records, and boundary deviation direction. Taking the current welding current of 165A, welding voltage of 25V, welding speed of 200mm / min, and wire feed speed of 5.8m / min as an example, the welding power is within the commonly used range for small to medium arc welding processes, and the welding speed is also suitable for multi-layer, multi-pass, or local temperature field control scenarios with a plate thickness of 12mm. When the soaking status code corresponds to the expansion of the first side and the heat accumulation on the first side continues, the candidate process parameters may include adjusting the welding current to 155A, adjusting the welding speed to 210mm / min, adjusting the wire feed speed to 5.5m / min, and slightly adjusting the welding torch oscillation center to the second side by 0.4mm. The above adjustments reduce the welding current, increase the welding speed, decrease the wire feed speed, and reduce the heat source effect on the outward expansion side, aligning with the direction of the return from the first side boundary towards the weld centerline. If another candidate process parameter adjusts the welding current to 175A or the welding speed to 190mm / min, this candidate process parameter will increase the overall heat input or prolong the heat source effect time, which is opposite to the boundary convergence direction required for the first side expansion, and should be deleted. The candidate process parameters after screening need to be compared with the boundary deviation data of the subsequent adjacent segment number 19; if the first side expansion already exists in segment number 19, the target process parameter with a gentle effect and continuous execution is allowed to be selected; if the boundary of segment number 19 is normal, the target process parameter is limited to execution only within the effective range of segment number 18. This process can reduce the impact of the current segment adjustment on the boundary state of the subsequent adjacent segments.

[0018] In S4, after the target process parameters are executed, the feedback is not based on whether a single temperature value decreases. Instead, the width of the heat-affected zone on both sides after execution is reacquired after a response waiting period and compared with the theoretical response data. The theoretical response data is obtained from the boundary deviation data before execution, the target process parameters, and historical valid execution records, including the direction of boundary change, the response waiting period, and the expected return segment. Taking a first-side heat-affected zone width of 10.4 mm, a first-side dynamic boundary data outer boundary of 9.1 mm, and a first-side offset distance of 1.3 mm before execution as an example, when the target process parameter is to reduce the effect of the first-side heat source, the boundary change direction in the theoretical response data is that the boundary of the first-side heat-affected zone returns towards the weld centerline. Historical valid execution records show that similar target process parameters enter a stable response after 12 control cycles, with a control cycle of 50 ms, therefore the response waiting time is 600 ms. Compared to the local welding temperature field change of a 12 mm thick plate, a waiting time of 600 ms can cover the initial response of power readback, travel speed change, and heat conduction hysteresis, making it more suitable as the boundary response sampling time than 300 ms. If the width of the first-side heat-affected zone becomes 9.6mm after execution, and the expected return range is 9.3mm to 9.8mm, the execution residual data is recorded as convergence in the same direction; if the width of the first-side heat-affected zone becomes 10.3mm after execution, and the target process parameters have been correctly read back, it is recorded as insufficient response; if the width of the first-side heat-affected zone becomes 10.8mm after execution, it is recorded as divergence in the opposite direction. This processing enables subsequent control actions to maintain, enhance, pause, or cancel the corresponding actions based on the execution results.

[0019] In summary, this implementation method forms a continuous processing chain through weld seam control sections, temperature mapping tables, unfolded temperature points, width of the double-sided heat-affected zone, dynamic boundary data, section-level backtracking data, heat soaking status codes, target process parameters, and execution residual data. Each data item corresponds to a clear input and output. The current heat-affected zone boundary point is first determined by the current temperature distribution and then compared with the dynamic boundary data, avoiding the use of dynamic boundary data as the basis for generating the current boundary point. Candidate process parameters are not generated by uninterpretable processes, but are jointly determined by the current welding process parameters, the minimum adjustment unit of the equipment, historical valid execution records, and heat input direction screening rules. Execution residual data does not depend on changes in a single temperature value, but compares the relationship between the theoretical response data and the width of the double-sided heat-affected zone after execution, making subsequent control actions and dynamic boundary data corrections verifiable.

[0020] like Figure 2 As shown, in one specific embodiment, S1 includes: S11. Read the tank geometry data and weld path data, and determine the division direction of the weld control section according to the weld type; S12. Project the temperature measurement position onto the start and end range of the weld control section, and determine the weld centerline side based on the weld centerline to generate a temperature measurement mapping table. S13. Read the temperature measurement value of the current sampling period according to the temperature measurement mapping table, and generate the unfolded temperature point with segment number, side, unfolded position and confidence mark.

[0021] In this embodiment, it should be noted that in S11, the tank geometry data is used to determine the weld surface position and segmentable range, while the weld path data is used to determine the weld type, weld centerline, and welding direction. The tank geometry data may include a tank radius of 1800mm, a tank plate thickness of 12mm, a tank axial length of 6000mm, and the tank support position. The weld path data may include the type of weld (circumferential or longitudinal), the weld start position, the weld end position, and the welding direction. When the weld type is a circumferential weld, the weld control segment is divided along the circumferential arc length of the tank; when the weld type is a longitudinal weld, the weld control segment is divided along the axial length of the tank. The length of the weld control segment is determined jointly based on the moving distance of the welding heat source within a stable control time, the length of the temperature measurement coverage area, and historical boundary change records. When taking values, first read a limited number of historical welding records under the same tank plate thickness, weld type, and welding posture. Then, calculate the minimum spatial length of the heat-affected zone boundary that maintains a distinguishable change within a continuous control cycle, ensuring that each weld control segment includes at least a temperature measurement point near the centerline, a first-side temperature measurement point, and a second-side temperature measurement point. If the welding speed is 200 mm / min, the control cycle is 50 ms, the effective length of the temperature measurement coverage area in the circumferential direction is approximately 250 mm, and the boundary change can form a stable comparison within the range of 200 mm to 260 mm, then the commonly used segment length can be taken as 235 mm. For cases where the total length of the circumferential weld cannot be divided evenly by the commonly used segment length, the last weld control segment is recorded as a margin segment. In circumferential weld welding, the weld control segments have a closed relationship, but segment-level revisits only use the preceding adjacent weld control segments that have completed sampling and formed the width of the double-sided heat-affected zone. If the current weld control segment is the starting segment, and the preceding adjacent weld control segment in its closing direction has not yet formed an effective double-sided heat-affected zone width, then the preceding adjacent segment number record is empty or there is no preceding effective segment, and no segment-level back-visit data is generated; when the circumferential weld completes one revolution and enters the closed area, the preceding adjacent segment number points to the adjacent weld control segment that has formed an effective record according to the welding travel direction.

[0022] In S12, when projecting the temperature measurement position, the position is first converted to the tank surface coordinates, and then its corresponding circumferential arc length or axial position is determined along the weld centerline. When the weld type is a circumferential weld, the circumferential arc length corresponding to the temperature measurement position determines its weld control segment. When the weld type is a longitudinal weld, the axial position corresponding to the temperature measurement position determines its weld control segment. Then, the vertical distance from the temperature measurement position to the weld centerline determines the unfolding position. If the temperature measurement position falls near the boundary of two adjacent weld control segments, it is preferentially assigned to the weld control segment with the closer center distance, and the boundary proximity mark is recorded in the temperature measurement mapping table. The side division rule of the weld centerline is determined according to the welding direction and the local unfolding coordinates of the weld. Specifically, the weld centerline is first used as the baseline in the unfolding coordinates, and then the welding direction is taken as the positive direction. The left side or the preset normal side is fixedly defined as the first side, and the opposite side is fixedly defined as the second side. The preset width range of the area near the centerline is determined based on the weld bevel width, temperature measurement point positioning error, and weld centerline tracking error. Taking a bevel width of 4mm, a temperature measurement point positioning error of 0.5mm, and a weld centerline tracking error of 0.5mm as an example, a 2.5mm range on each side of the weld centerline can be recorded as the area near the centerline. Temperature measurement points exceeding this range but located on the first side are recorded as the first side, and those exceeding this range but located on the second side are recorded as the second side. The temperature measurement mapping table must include at least the temperature measurement point number, the corresponding weld control segment, the unfolding position, the weld centerline side, and the effective coverage period to ensure that each subsequent temperature measurement value can be written to the corresponding segment number and corresponding side.

[0023] In S13, the temperature value of the current sampling cycle is converted into the unfolded temperature point through a temperature mapping table. The unfolded temperature point includes the corresponding weld control section, weld centerline side, unfolded position, temperature value, sampling time, and temperature point reliability marker. The temperature measurement range is determined based on the calibration record of the actual temperature measuring equipment and the temperature measurement requirements of the current welding material. When taking the value, first read the factory-calibrated range and the field calibration range of the temperature measuring equipment, and then combine them with the current welding material, plate thickness, and process temperature range to determine the effective temperature measurement range; if the field calibration range is smaller than the factory-calibrated range, the field calibration range is used as the temperature measurement range. Taking the factory-calibrated range of the infrared temperature measuring equipment as 0℃ to 1400℃ and the stable range after field calibration as 0℃ to 1200℃ as an example, 0℃ to 1200℃ is used as the temperature measurement range. The temperature point reliability marker is generated within the sliding sampling window. In the current control cycle, the temperature point that has entered the sliding sampling window is read first, and then the existence of a valid temperature value in the previous and next sampling cycles of that temperature point is checked. For a temperature point generated in the latest sampling cycle, if the next sampling cycle has not yet been formed, it is recorded as a temperature point to be confirmed and is not used as a separate basis for the boundary point of the heat-affected zone. It is then updated to the first reliable temperature point, the second reliable temperature point, or an invalid temperature point when the next sampling cycle arrives. In this way, real-time control does not need to wait for all welding to be completed, nor does it use data from the next sampling cycle that has not yet been formed.

[0024] The effective coverage period is determined based on the time relationship between the temperature measurement point's coverage area and the welding heat source passing through the corresponding weld control section. When taking the value, first read the temperature measurement coverage area of ​​the temperature measurement point, then read the welding speed and the start and end range of the weld control section. Determine the sampling period for the welding heat source entering the temperature measurement coverage area, passing through the center of the temperature measurement coverage area, and leaving the temperature measurement coverage area. Write the continuous sampling period that effectively reflects the temperature field change of the weld control section into the temperature measurement mapping table. For example, with a welding speed of 200 mm / min, a control period of 50 ms, and a temperature measurement point covering 20 mm along the weld direction, it takes approximately 6 seconds for the welding heat source to pass through the coverage area, corresponding to approximately 120 control periods. The temperature measurement point can be recorded as the effective coverage period within these 120 control periods. Adjacent missing sampling points are determined jointly based on adjacent temperature measurement points at the deployment position and adjacent sampling periods in time. Temperature points outside the boundary refer to temperature points whose deployment position is already outside the effective area of ​​their respective weld control section, or located outside the edge of the temperature measurement coverage area and cannot stably reflect the temperature field of the weld control section. For example, if the circumferential range of segment number 18 is 3995mm to 4230mm, and the circumferential position of a certain temperature measurement point after unfolding is 4238mm, this temperature measurement point should not be written into the first reliable temperature point of segment number 18, but should be assigned to the adjacent segment or recorded as a temperature point outside the boundary.

[0025] like Figure 3 and Figure 4 As shown, in one specific embodiment, S2 includes: S21. Within the same weld control section, read the temperature points on the first and second sides along the direction away from the weld centerline to obtain the width of the double-sided heat-affected zone. S22. Compare the width of the heat-affected zone on both sides of the current weld control section with the width on the same side and the corresponding dynamic boundary data of the adjacent weld control section in the previous sequence to obtain boundary deviation data and section-level back visit data. S23. Generate a uniform heating status code from boundary deviation data and segment-level return visit data.

[0026] In this embodiment, it should be noted that in S21, the temperature points are first divided into first-side temperature points, second-side temperature points, and temperature points near the centerline according to the weld centerline within the same weld control section. Then, the first-side and second-side temperature points are arranged in directions away from the weld centerline. Invalid temperature points are not included in the determination of heat-affected zone boundary points, and the second reliable temperature point is only used as an auxiliary comparison point and not as a separate heat-affected zone boundary point. The heat-affected zone edge temperature records in historical qualified welds are determined based on the temperature measurement records of a limited number of qualified welds and the post-weld quality confirmation results. When taking values, first, historical weld batches consistent with the current tank material, tank plate thickness, weld type, welding posture, and temperature mapping table version are selected. Then, batches with qualified post-weld appearance inspection, non-destructive testing, or process release records are selected. The temperature values ​​corresponding to the heat-affected zone edge positions in these batches are read, and data with missing sampling, temperature exceeding the measurement range, and execution failure cycles are deleted. Taking the temperature records of the heat-affected zone (HAZ) edge of five qualified welds as examples (735℃, 742℃, 739℃, 746℃, and 738℃ respectively), 735℃ to 746℃ can be used as the historical HAZ edge temperature records for qualified welds under the current material and plate thickness. If the temperature values ​​at positions 6mm, 8mm, 10.4mm, and 12mm perpendicular to the weld direction on the first side are 812℃, 780℃, 742℃, and 690℃ respectively, then the 12mm position is already below the lower limit of the historical HAZ edge temperature records for qualified welds, while the 10.4mm position is within that temperature record. Therefore, the 10.4mm position is recorded as the boundary point of the HAZ on the first side.

[0027] In S22, boundary deviation data is generated based on the relationship between the current heat-affected zone width and the corresponding dynamic boundary data. Specifically, it first determines whether the width of the first-side heat-affected zone is within the first-side dynamic boundary data, and then determines whether the width of the second-side heat-affected zone is within the second-side dynamic boundary data. If the width of the corresponding-side heat-affected zone is greater than the outer boundary of the dynamic boundary data, the deviation direction is recorded as outward expansion; if the width of the corresponding-side heat-affected zone is less than the inner boundary of the dynamic boundary data, the deviation direction is recorded as insufficient; if the width of the corresponding-side heat-affected zone is within the dynamic boundary data, the deviation direction is recorded as normal. The offset distance is recorded according to the length difference between the current heat-affected zone width and the nearest boundary of the corresponding-side dynamic boundary. If the width of the first-side heat-affected zone is 10.4mm, and the first-side dynamic boundary data is 8.8mm to 9.1mm, then the first-side deviation direction is outward expansion, and the first-side offset distance is 1.3mm; if the width of the second-side heat-affected zone is 8.9mm, and the second-side dynamic boundary data is 8.7mm to 9.2mm, then the second-side deviation direction is normal. Subsequently, the first and second side deviation directions of the preceding adjacent segment number 17 are read. If the width of the first side heat-affected zone of segment number 17 is 10.0 mm, its first side deviation direction is also outward expansion, and the second side deviation direction of segment number 17 is normal, then the segment-level back-visit data between the current segment number 18 and the preceding adjacent segment number 17 is recorded as the first side heat accumulation continuation. The segment-level back-visit data refers to the boundary deviation continuation relationship between the current weld control segment and the preceding adjacent weld control segment. Its corresponding general technical object is the boundary comparison result of adjacent weld segments. The input data includes the boundary deviation data of the current weld control segment, the boundary deviation data of the preceding adjacent weld control segment, and the welding direction. The generation method is to compare whether the same side deviation direction appears continuously between adjacent segments. The output results include the first side heat accumulation continuation, the second side heat accumulation continuation, the overall heat accumulation continuation, the current segment local offset, and the heat history interruption. The type is an enumeration identifier, the usage stage is the generation of the heat soaking status code, and the technical function is to distinguish between the local deviation of the current weld control segment and the heat accumulation that continues along the welding direction.

[0028] In S23, the heat homogenization status code is generated according to a finite enumeration rule. Boundary deviation data is first converted into boundary deviation direction code, bilateral balance code, and return persistence code. Segment-level backtracking data is converted into thermal history backtracking code. Then, these are combined in the order of boundary deviation direction code, thermal history backtracking code, bilateral balance code, and return persistence code to form the heat homogenization status code. The boundary deviation direction code is used to indicate overall expansion, overall insufficiency, first-side expansion, second-side expansion, first-side insufficiency, second-side insufficiency, or normal boundary. The thermal history backtracking code is used to indicate continuation of first-side thermal accumulation, continuation of second-side thermal accumulation, continuation of overall thermal accumulation, local offset of the current segment, or interruption of thermal history. The bilateral balance code is used to indicate bilateral balance or bilateral imbalance. The return persistence code is used to indicate that the segment has returned, is returning, has not returned, or is continuously increasing. Taking segment number 18 as an example, the first side deviation direction is outward expansion, the second side deviation direction is normal, the segment-level back-visit data shows the continuation of heat accumulation on the first side, the outward expansion on the first side has occurred continuously for 5 control cycles and the number of returned samples is 0, therefore the boundary deviation direction code is the outward expansion on the first side, the thermal history back-visit code is the continuation of heat accumulation on the first side, the bilateral balance code is the bilateral imbalance, and the return persistence code is no return. This heat equalization status code corresponds to single-sided boundary offset control, and subsequent candidate process parameters need to reduce the effect of the heat source on the first side and cannot increase the effect of the heat source on the first side. The number of persistence cycles and the number of returned samples are obtained from the statistics of continuous cycle records. If the first side of segment number 18 is greater than the outer side of the dynamic boundary of the first side in the last 5 control cycles, and no valid sample falls within the range of 8.8mm to 9.1mm, then the number of persistence cycles is 5 and the number of returned samples is 0.

[0029] Dynamic boundary data is generated from historical set of expanded temperature points, historical widths of the heat-affected zone on both sides, and historical execution records. Effective stable windows are selected based on data integrity records and window validity records. Specifically, the number of control cycles within each candidate historical window is first fixed to the same number, then the number of missing temperature points, invalid temperature points, boundary direction reversals, and execution failure markers within that window are counted. The number of missing temperature points refers to the number of times a valid temperature measurement was not obtained at a location where a temperature value should have been acquired; the number of invalid temperature points refers to the number of times a temperature measurement value exceeded the range, the coordinates could not be mapped, or it was marked as an invalid temperature point; the number of boundary direction reversals refers to the number of times the boundary of the heat-affected zone on the same side changed from outward expansion to inward contraction or vice versa within adjacent control cycles; execution failure markers indicate that the target process parameters were not correctly read back by the welding control system or the actuator returned a failure status. The allowable number of integrity points, invalid points, and reversals are all obtained from the stable windows in historical qualified batches. For example, each candidate historical window contains 60 control cycles, and each control cycle requires 12 valid temperature measurement points, for a total of 720 temperature points to be collected. If the width of the bilateral heat-affected zone can still be stably identified when the number of missing temperature points shown in the historical qualified batches does not exceed 3, then the allowable number of integrity points is 3. If the width of the bilateral heat-affected zone can still be obtained when the number of invalid temperature points is 2, but the first boundary point is continuously missing when the number of invalid temperature points is 5, then the allowable number of invalid points is 2. If the number of boundary direction reversals in the historical qualified window is usually 1 to 3 times, while it reaches more than 10 times in the abnormal window, then the allowable number of reversals is 3. When the width of the first side heat-affected zone recorded in the effective stable window is 8.8mm, 8.9mm, 9.0mm, 9.1mm, and 8.9mm, the dynamic boundary data of the first side is 8.8mm to 9.1mm. The dynamic boundary data is used for boundary deviation judgment and does not participate in the initial identification of the current heat-affected zone boundary point.

[0030] like Figure 5 and Figure 6 As shown, in one specific embodiment, S3 includes: S31. Obtain candidate process parameters from the heat soaking status code and the current welding process parameters; S32. Screen out candidate process parameters with opposite directions according to the heat input direction defined by the heat homogenization status code; S33. Combine the boundary deviation data of the subsequent adjacent weld control sections to obtain the target process parameters.

[0031] In this embodiment, it should be noted that in S31, the generation of candidate process parameters is based on the current welding process parameters, the minimum adjustment unit of the equipment, historical effective execution records, and the direction of boundary deviation. The current welding process parameters may include welding current, welding voltage, welding speed, wire feed speed, welding torch oscillation center, and heat source action time. The minimum adjustment unit of the equipment is determined based on the control interface parameters of the welding power source, the traveling mechanism, the wire feed mechanism, and the welding torch oscillation mechanism. When determining the value, first read the minimum change allowed by the equipment control system, and then confirm the actual executable minimum change in conjunction with the equipment field calibration record; if the change allowed by the control interface is less than the change that can be stably executed on-site, then the change that can be stably executed on-site is taken as the minimum adjustment unit of the equipment. Taking the welding power source allowing the current command to be issued at 1A, but the field calibration showing that the adjustment of less than 5A is unstable in the readback value as an example, the minimum adjustment unit of the welding current can be set to 5A; if the traveling mechanism can stably execute a speed change of 10mm / min, then the minimum adjustment unit of the welding speed is set to 10mm / min. When the first side offset distance of segment 18 is 1.3mm, and historical valid execution records show that when the first side expands outward by 1.0mm to 1.5mm, reducing the welding current by 10A and increasing the welding speed by 10mm / min can achieve convergence in the same direction after 12 control cycles. Therefore, the current candidate process parameters that can be generated are: adjusting the welding current from 165A to 155A and the welding speed from 200mm / min to 210mm / min. The adjusted arc power and heat input per unit length are both reduced, which is consistent with the control direction of reducing heat input when the boundary expands outward.

[0032] In S32, heat input direction screening is determined based on the influence of process parameters on the heat input direction. If the heat soaking status code corresponds to a reduction in heat input, then the welding current and welding voltage in the candidate process parameters must not be higher than the current value, the welding speed must not be lower than the current value, the wire feed speed must not be higher than the current value, and the heat source action time must not be extended. If the heat soaking status code corresponds to an increase in heat input, then the welding current and welding voltage in the candidate process parameters must not be lower than the current value, the welding speed must not be higher than the current value, the wire feed speed must not be lower than the current value, and the heat source action time must not be shortened. If the heat soaking status code corresponds to single-sided boundary offset control, then it is also necessary to determine whether the welding torch oscillation center and heat source action time reduce the effect on the outward expansion side or increase the effect on the insufficient side. Taking the first-sided outward expansion as an example, if the candidate process parameter moves the welding torch oscillation center to the first side by 0.4mm, it will increase the heat source effect on the first side and will not pass the screening rule. If the candidate process parameter moves the welding torch oscillation center to the second side by 0.4mm and does not increase the overall heat input, then the candidate process parameter is retained. If the current welding current is 165A, and a candidate process parameter increases the welding current to 175A, this candidate process parameter is contrary to the direction of reducing heat input and should be deleted. If the current welding speed is 200mm / min, and a candidate process parameter decreases the welding speed to 190mm / min, this will increase the heat source residence time and should also be deleted. This process only compares the relationship between the direction of process parameter change and the direction of heat input, and does not directly add data of different dimensions.

[0033] In S33, the boundary deviation data of subsequent adjacent weld seam control segments are used to limit the execution range and adjustment amplitude of candidate process parameters after screening. Specifically, the first and second side deviation directions of the subsequent adjacent weld seam control segments are read first, and then it is determined whether the heat input direction corresponding to the candidate process parameter is opposite to the deviation convergence direction of the subsequent adjacent weld seam control segments. If the current weld seam control segment is insufficient overall, while the subsequent adjacent weld seam control segments have already expanded overall, then candidate process parameters that significantly increase heat input are not adopted; instead, candidate process parameters with a smaller amplitude or those executed only within the current weld seam control segment are selected. If both the current weld seam control segment and the subsequent adjacent weld seam control segments have expanded first side, then candidate process parameters that reduce the effect of the first side heat source can be selected, and continuous execution within two adjacent weld seam control segments is allowed. Taking the expansion of the first side of segment 18 and the expansion of the first side of segment 19 as an example, the target process parameters can be selected by slightly adjusting the welding torch oscillation center to the second side by 0.4mm, and coordinating with a 10A reduction in welding current. If the boundary of segment 19 is normal, the target process parameters will only be executed within the effective range of segment 18. The target process parameters are the execution parameters obtained after filtering by heat input direction and limiting the deviation of adjacent segment boundaries. The output results can include the target welding current, target welding voltage, target welding speed, target wire feed speed, target welding torch oscillation center, and target heat source action time.

[0034] In another specific implementation, candidate process parameters can be divided into overall heat input candidates and single-sided effect candidates. Overall heat input candidates include welding current, welding voltage, welding speed, wire feed speed, and heat source application time, primarily used for overall expansion or overall deficiency. Single-sided effect candidates include welding torch oscillation center, welding torch lateral position, and heat source application time allocation on both sides, primarily used for first-sided expansion, second-sided expansion, first-sided deficiency, or second-sided deficiency. If the heat soaking status code indicates overall expansion and balanced application on both sides, candidates that reduce overall heat input are prioritized. If the heat soaking status code indicates first-sided expansion and normal application on the second side, single-sided effect candidates that reduce the application of heat sources on the first side are prioritized, avoiding direct reduction of overall heat input that could cause second-sided deficiency. Through this candidate grouping method, the generation of target process parameters can maintain consistency with the boundary deviation type corresponding to the heat soaking status code.

[0035] like Figure 7 and Figure 8 As shown, in one specific embodiment, S4 includes: S41. The theoretical response data is obtained from the pre-execution boundary deviation data, target process parameters, and historical valid execution records; S42. After the response waiting period, the width of the heat-affected zone on both sides after execution is re-acquired, and the width of the heat-affected zone on both sides after execution is compared with the theoretical response data to obtain the execution residual data. S43. Correct subsequent control actions and dynamic boundary data using the execution residual data.

[0036] In this embodiment, it should be noted that in S41, the theoretical response data is obtained from the pre-execution boundary deviation data, the target process parameters, and historical valid execution records. The theoretical response data includes the boundary change direction, the response waiting period, and the expected return segment. The theoretical response data does not predict the entire temperature field, but rather records which direction the heat-affected zone boundary should change after the execution of the current target process parameters, how many control cycles to wait before reading the post-execution boundary, and the length range the post-execution heat-affected zone width should enter. The response waiting period is determined based on the number of control cycles required for a stable boundary response to be generated by similar target process parameters in the historical valid execution records. When determining the value, historical execution records consistent with the current tank plate thickness, weld type, welding posture, boundary deviation direction, and process parameter adjustment type are first selected. Then, the number of cycles in which the heat-affected zone width first continuously changes along the theoretical response direction after the execution of the target process parameters is read, and records where the target process parameters were not correctly read back or where the temperature measurement data is invalid are excluded. Taking a historical effective execution record where the welding current is reduced by 10A and the welding speed is increased by 10mm / min, and the width of the first-side heat-affected zone typically begins to stabilize and return towards the weld centerline after the 12th control cycle, as an example, if the control cycle is 50ms, then the response waiting cycle is 12 control cycles, corresponding to a waiting time of 600ms. The expected return range is determined by the stable return range after the execution of similar target process parameters in the historical effective execution record. If the historical effective execution record shows that the width of the first-side heat-affected zone typically returns to 9.3mm to 9.8mm after the response waiting cycle, then 9.3mm to 9.8mm can be taken as the expected return range for the current similar control action.

[0037] In S42, the execution residual data is classified according to the relationship between the direction of boundary change after execution and the direction of boundary change in the theoretical response data. Specifically, the offset distance in the boundary deviation data before execution is read first, and then the width of the heat-affected zone on both sides after execution is read after the response waiting period. It is then determined whether the boundary of the heat-affected zone after execution moves in the direction required by the theoretical response data. If the theoretical response data requires the first side boundary to return towards the weld centerline, and the width of the first side heat-affected zone changes from 10.4mm to 9.6mm after execution, and 9.6mm is within the expected return range of 9.3mm to 9.8mm, then the execution residual data is recorded as convergence in the same direction. If the width of the first side heat-affected zone changes from 10.4mm to 10.3mm after execution, although the boundary returns slightly, it does not enter the expected return range, then the execution residual data is recorded as insufficient response. If the width of the first side heat-affected zone changes from 10.4mm to 10.8mm after execution, then the execution residual data is recorded as divergence in the opposite direction. Whether the target process parameters are correctly read back is judged based on the consistency between the actual execution parameters returned by the welding control system and the target process parameters. Specifically, after issuing the target process parameters, the actual parameter values ​​returned by the welding power supply, wire feeding mechanism, traveling mechanism, and welding torch oscillation mechanism are read and compared item by item with the target process parameters. The readback error range is determined based on the control accuracy of the welding equipment, the response accuracy of the transmission mechanism, and the on-site calibration records. Taking a target welding current of 155A as an example, if the on-site calibration shows a readback value between 154A and 156A, which is considered normal execution, then the current readback error range can be taken as 1A; if the target welding speed is 210mm / min, and the readback value is stable between 208mm / min and 212mm / min, then the speed readback error range can be taken as 2mm / min.

[0038] In S43, the execution residual data corrects subsequent control actions and dynamic boundary data. If the execution residual data shows convergence in the same direction, it indicates that the target process parameters have caused the heat-affected zone width to return along the theoretical response direction. Subsequent control actions maintain the current direction and reduce the adjustment amplitude to avoid excessive return caused by continuing to adjust with the same amplitude. If the execution residual data shows insufficient response, and the target process parameters have been correctly read back, it indicates that the control direction is consistent with the boundary convergence direction but the adjustment amplitude is insufficient. Subsequent control actions maintain the current direction and increase the adjustment amplitude. If the execution residual data shows insufficient response, and the target process parameters have not been correctly read back, then similar control actions are paused and execution link verification is initiated to avoid mistaking the failure of the execution link for insufficient heat input adjustment. If the execution residual data shows divergence in the opposite direction, then control actions in the same direction are canceled and candidate process parameters are regenerated. The number of multiple consecutive control cycles is determined based on the control cycle length, temperature sampling stability, and historical execution response lag time. If the control cycle is 50ms, historical records show that the width of the heat-affected zone may temporarily change due to temperature fluctuations within a single control cycle, but a stable trend is usually observed when it increases in the same direction for three consecutive control cycles. Therefore, three consecutive control cycles can be taken. Taking the width of the heat-affected zone on the first side continuously increasing from 10.4mm to 10.6mm and then to 10.8mm as an example, this can be recorded as the execution residual data increasing in the same direction for three consecutive control cycles.

[0039] Dynamic boundary data is only corrected within the subsequent effective stable window and is not directly modified during the period of anomaly persistence. Post-weld quality confirmation results are obtained based on at least one of the following: post-weld visual inspection, weld size inspection, non-destructive testing, or process release records. When taking values, first read the post-weld quality record of the corresponding weld control section or the corresponding weld batch, and then confirm whether there are any records of incomplete fusion, undercut, cracks, excessive deformation, or non-destructive testing failures. If no quality problems exceeding process requirements are found, the post-weld quality confirmation result of that weld batch or the corresponding weld control section is recorded as qualified. If quality problems exist, the batch data is not allowed to be used for dynamic boundary data outward or inward movement. The boundary version number is used to record the order of use and correction of dynamic boundary data. After each generation or correction of dynamic boundary data, a new boundary version number is written, and this boundary version number is associated with and saved with subsequently generated boundary deviation data, section-level backtracking data, soaking status codes, target process parameters, and execution residual data. If the dynamic boundary data on the first side is corrected from 8.8mm to 9.1mm to 9.2mm to 9.4mm, the corrected dynamic boundary data is written with a new boundary version number. Data already generated in the current control cycle is not reverted, and the new dynamic boundary data is used starting from the next control cycle. This process ensures consistency in the comparison criteria within the same control cycle.

[0040] like Figure 9As shown, this invention also provides a system for controlling the local homogenization temperature field of tank welds, including: a weld control segment generation module, a temperature measurement mapping module, a unfolded temperature point generation module, a double-sided heat-affected zone boundary identification module, a segment-level comparison module, a homogenization status code generation module, a target process parameter generation module, a control execution module, an execution residual generation module, a correction module, and a historical time series database. The weld control segment generation module is used to divide the tank circumferential weld into multiple weld control segments according to the circumferential arc length or the tank longitudinal weld according to the axial length. The temperature measurement mapping module is used to write the temperature measurement point number, unfolded position, and weld centerline side into a temperature measurement mapping table. The unfolded temperature point generation module is used to convert the temperature measurement value of the current sampling period into an unfolded temperature point with segment number, side, unfolded position, and reliable marker according to the temperature measurement mapping table. The double-sided heat-affected zone boundary identification module is used to read the first side temperature point and the second side temperature point along the direction away from the weld centerline within the same weld control segment to obtain the width of the double-sided heat-affected zone. The segment-level comparison module compares the width of the heat-affected zone (HAZ) on both sides of the current weld control segment with the width on the same side and the corresponding dynamic boundary data of the preceding adjacent weld control segment to obtain boundary deviation data and segment-level backtracking data. The heat-aware status code generation module generates a heat-aware status code from the boundary deviation data and segment-level backtracking data. The target process parameter generation module filters out candidate process parameters with opposite directions according to the heat input direction defined by the heat-aware status code, and combines this with the boundary deviation data of subsequent adjacent weld control segments to obtain the target process parameters. The control execution module executes the target process parameters. The execution residual generation module compares the width of the HAZ after execution with the theoretical response data to obtain execution residual data. The correction module corrects subsequent control actions and dynamic boundary data based on the execution residual data. The historical time-series database stores weld control segments, temperature mapping tables, unfolded temperature points, width of the HAZ on both sides, boundary deviation data, segment-level backtracking data, heat-aware status codes, target process parameters, execution residual data, and dynamic boundary data.

[0041] In a specific operational scenario, the historical time-series database stores historical records corresponding to the current tank plate thickness of 12mm, circumferential weld type, and the same welding posture. There are four candidate historical windows, each containing 60 control cycles, with each control cycle lasting 50ms. Window A has 0 missing temperature points, 1 invalid temperature point, and 2 boundary direction flips; execution failure is marked as non-existent. Window B has 1 missing temperature point, 1 invalid temperature point, and 3 boundary direction flips; execution failure is marked as non-existent. Window C has 8 missing temperature points, 6 invalid temperature points, and 11 boundary direction flips; execution failure is marked as present. Window D has 0 missing temperature points, 0 invalid temperature points, and 2 boundary direction flips; execution failure is marked as non-existent. When the allowed number of integrity values ​​is 3, the allowed number of invalid values ​​is 2, and the allowed number of flips is 3, windows A, B, and D are considered valid stable windows, while window C is excluded. The widths of the first heat-affected zone within the effective stabilization window were recorded as 8.8 mm, 8.9 mm, 9.0 mm, 9.1 mm, and 8.9 mm, and the widths of the second heat-affected zone were recorded as 8.7 mm, 8.8 mm, 9.0 mm, 9.2 mm, and 8.9 mm. Thus, the dynamic boundary data for the first side ranged from 8.8 mm to 9.1 mm, and the dynamic boundary data for the second side ranged from 8.7 mm to 9.2 mm.

[0042] The following example, using specific parameters and data, illustrates the method for controlling the local isothermal temperature field of a tank weld and the system's operation. In a scenario involving circumferential welding of a carbon steel tank, the tank radius is 1800mm, the tank plate thickness is 12mm, and the circumferential length of the circumferential weld is 1.131 × 10⁻⁶ mm. 4The welding travel direction is clockwise, and the control cycle is 50ms. The circumferential weld is divided into 48 weld control segments according to the circumferential arc length. The first 47 weld control segments are 235mm long, and the 48th weld control segment is a margin segment with a length of approximately 265mm. The circumferential starting point of segment number 18 is 3995mm, the circumferential ending point is 4230mm, and the center position of the segment is 4112.5mm. In S1, the curved surface temperature measurement position of temperature measurement point 035 is projected onto segment 18, with an unfolded position of 4100mm circumferential arc length and 6mm perpendicular to the weld direction, and the side is designated as the first side; the unfolded position of temperature measurement point 036 is 4105mm circumferential arc length and 8mm perpendicular to the weld direction, and the side is designated as the first side; the unfolded position of temperature measurement point 037 is 4110mm circumferential arc length and 10.4mm perpendicular to the weld direction, and the side is designated as the first side; the unfolded position of temperature measurement point 038 is 4115mm circumferential arc length and 12mm perpendicular to the weld direction, and the side is designated as the first side. At sampling time 18.50s, the temperature measurement value of temperature measurement point 035 is 812℃, the temperature measurement value of temperature measurement point 036 is 780℃, the temperature measurement value of temperature measurement point 037 is 742℃, and the temperature measurement value of temperature measurement point 038 is 690℃. Since the temperature measurement range is from 0℃ to 1200℃, and the above temperature measurement points are sampled continuously within the sliding sampling window, temperature measurement point numbers 035, 036, and 037 are all marked as the first reliable temperature points; if the adjacent temperature measurement point outside temperature measurement point number 038 is missing, then temperature measurement point number 038 can be used as the second reliable temperature point, and is not used as the boundary point of the heat-affected zone alone.

[0043] In S2, the edge temperature records of the heat-affected zone (HAZ) in historical qualified welds are determined by the edge temperature measurements of five qualified welds: 735℃, 742℃, 739℃, 746℃, and 738℃. Therefore, the edge temperature records range from 735℃ to 746℃. In segment 18, the temperature measurement at 10.4mm perpendicular to the weld direction is 742℃, which falls within the historical HAZ edge temperature records for qualified welds. The temperature measurement at 12mm perpendicular to the weld direction is 690℃, which is below the lower limit of the edge temperature record. Therefore, the 10.4mm position is taken as the boundary point of the HAZ on the first side, and the width of the HAZ on the first side is 10.4mm. The second side of segment 18 is obtained with the same rule, resulting in a HAZ width of 8.9mm. The effective stability window provides dynamic boundary data of 8.8mm to 9.1mm for the first side and 8.7mm to 9.2mm for the second side. Since the width of the first-side heat-affected zone (HAD) is 10.4 mm, which is greater than the outer boundary of the first-side dynamic boundary data (9.1 mm), the deviation direction of the first side is recorded as outward expansion. The offset distance of the first side is obtained by the difference between 10.4 mm and 9.1 mm, which is 1.3 mm. The width of the second-side HAD is 8.9 mm, which is within the range of 8.7 mm to 9.2 mm. The deviation direction of the second side is recorded as normal. The width of the first-side HAD of the adjacent segment 17 is 10.0 mm, and the width of the second-side HAD is 8.8 mm. Therefore, the deviation direction of the first side of segment 17 is also outward expansion, and the deviation direction of the second side is normal. The deviation directions of the first side of both segment 18 and segment 17 are outward expansion, and the deviation directions of the second side are both normal. Therefore, the segment-level back-visit data is recorded as the continuation of the first-side heat accumulation.

[0044] In S2, the boundary deviation data also includes the number of duration cycles and the number of return samples. In the last five control cycles, the width of the first-side heat-affected zone of segment 18 was 9.6mm, 9.8mm, 10.0mm, 10.2mm, and 10.4mm, respectively, all greater than the outer boundary of the first-side dynamic boundary data (9.1mm), therefore the number of duration cycles is 5. In none of these five control cycles did the width of the first-side heat-affected zone return to the range of 8.8mm to 9.1mm, therefore the number of return samples is 0. The boundary deviation direction code is for the first-side outward expansion, the thermal history return code is for the continuation of heat accumulation on the first side, the bilateral balance code is for bilateral imbalance, and the return duration code is for no return. These are combined in a fixed order to form the heat homogenization status code. The processing direction corresponding to this heat homogenization status code is to reduce the effect of the heat source on the first side, rather than simultaneously reducing the overall heat input on both sides by the same magnitude. The reason for this approach is that the width of the second-side heat-affected zone, at 8.9 mm, is already within the dynamic boundary data of 8.7 mm to 9.2 mm. If only an overall reduction in heat input is adopted, it may result in insufficient heat input on the second side. By using the heat homogenization status code, the expansion of the first side, the continuation of heat accumulation on the first side, the imbalance between the two sides, and the failure to return are all recorded together. Subsequent process parameters are then used to filter out candidates that can reduce the heat source effect on the first side and to delete candidates that will increase the heat source effect on the first side.

[0045] In S3, the current welding process parameters are: welding current 165A, welding voltage 25V, welding speed 200mm / min, and wire feed speed 5.8m / min. The minimum adjustment unit for the equipment is: welding current 5A, welding speed 10mm / min, wire feed speed 0.1m / min, and welding torch oscillation center 0.1mm. Historical valid execution records show that for a 1.0mm to 1.5mm outward expansion on the first side, reducing the welding current by 10A, increasing the welding speed by 10mm / min, decreasing the wire feed speed by 0.3m / min, and fine-tuning the welding torch oscillation center towards the second side by 0.4mm typically allows the boundary of the heat-affected zone on the first side to return to the weld centerline within the response waiting period. Therefore, candidate process parameters include: welding current 155A, welding voltage 25V, welding speed 210mm / min, wire feed speed 5.5m / min, and a 0.4mm fine-tuning of the welding torch oscillation center towards the second side. Simultaneously, another candidate process parameter generated, if it includes a welding current of 175A or a welding speed of 190mm / min, will increase heat input or prolong the heat source dwell time, which is opposite to the convergence direction of the first side expansion and is therefore deleted. Then, the boundary deviation data of the subsequent adjacent segment number 19 is read. If the width of the heat-affected zone on the first side of segment number 19 is 9.6mm, which is also greater than the outer boundary of the dynamic boundary data of the first side (9.1mm), then segment number 19 also has a first side expansion, allowing the target process parameter to reduce the effect of the first side heat source to be executed smoothly across adjacent segments; if the width of the heat-affected zone on the first side of segment number 19 is 9.0mm, then the target process parameter is limited to execution within segment number 18.

[0046] In S4, the width of the first-side heat-affected zone before execution is 10.4 mm, the outer boundary of the first-side dynamic boundary data is 9.1 mm, and the first-side offset distance is 1.3 mm. The target process parameters are welding current 155 A, welding voltage 25 V, welding speed 210 mm / min, wire feed speed 5.5 m / min, and a 0.4 mm fine adjustment of the welding torch oscillation center towards the second side. The boundary change direction in the theoretical response data is the return of the first-side heat-affected zone boundary towards the weld centerline. The response waiting period is 12 control cycles, and the control cycle is 50 ms, so the waiting time is 600 ms, and the expected return segment is 9.3 mm to 9.8 mm. After the target process parameters are issued, the welding current of 155 A, the welding speed of 210 mm / min, and the wire feed speed of 5.5 m / min are read back and found to be within the allowable readback error range, indicating that the target process parameters have been correctly read back. After 600ms, if the width of the first-side heat-affected zone after execution is 9.6mm, which is within the expected return range, the execution residual data is recorded as convergence in the same direction. Subsequent control actions maintain the direction of reducing the effect of the first-side heat source and reduce the adjustment amplitude, for example, the next current adjustment amplitude is reduced from 10A to 5A. If the width of the first-side heat-affected zone after execution is 10.3mm, and the readback parameters are correct, the execution residual data is recorded as insufficient response, and subsequent actions maintain the same direction and increase the adjustment amplitude. If the readback welding current is still 165A, similar control actions are paused and execution link verification is initiated.

[0047] In S4, reverse divergence and dynamic boundary data correction also need to be handled. If the width of the first-side heat-affected zone changes from 10.4mm to 10.8mm after execution, it indicates that the boundary has moved in the opposite direction to the boundary change direction in the theoretical response data. The execution residual data is recorded as reverse divergence, the same-direction control action is canceled, and candidate process parameters are regenerated. If, in the subsequent three consecutive control cycles, the width of the first-side heat-affected zone is 10.4mm, 10.6mm, and 10.8mm respectively, and the target process parameters are all correctly read back, it is recorded as continuous same-direction growth of the execution residual data, but the dynamic boundary data is not immediately modified. Only when the width of the first-side heat-affected zone remains stable within the new range for a long period in the subsequent effective stable window, and the post-weld visual inspection, non-destructive testing, and dimensional inspection are all qualified, is it allowed to correct the dynamic boundary data. For example, if the original dynamic boundary data for the first side was 8.8mm to 9.1mm, and the subsequent effective stabilization window shows that the width of the heat-affected zone on the first side is stable at 9.2mm to 9.4mm, and the post-weld quality confirmation result is qualified, then the dynamic boundary data for the first side will be shifted outward to 9.2mm to 9.4mm, and a new boundary version number will be written. Boundary deviation data, heat soaking status codes, and execution residual data already formed in the current control cycle will not be modified. The new dynamic boundary data will be used starting from the next control cycle. This process avoids temporary data during abnormal periods directly contaminating the dynamic boundary data and also ensures that boundary corrections correspond to the post-weld quality results.

[0048] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0049] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0050] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for intelligent control of the local heat equalization temperature field of tank welds, characterized in that, The methods include: Divide the tank circumferential seam into weld control sections according to the circumferential arc length or the tank longitudinal seam according to the axial length, and write the temperature measurement point number, unfolding position and weld center line side into the temperature measurement mapping table respectively. The temperature measurement value of the current sampling period is converted into an unfolded temperature point with segment number, side, unfolded position and confidence mark according to the temperature measurement mapping table; Within the same weld control section, temperature points on the first and second sides are read along the direction away from the weld centerline to obtain the width of the double-sided heat-affected zone. The width of the heat-affected zone on both sides of the current weld control section is compared with the width on the same side and the corresponding dynamic boundary data of the previous adjacent weld control section to obtain boundary deviation data and section-level back visit data. A heat-equalizing status code is generated from boundary deviation data and segment-level back-visit data. Candidate process parameters with opposite directions are screened out according to the heat input direction defined by the heat-equalizing status code. The target process parameters are obtained by combining the boundary deviation data of the subsequent adjacent weld control segments. After executing the target process parameters, the width of the heat-affected zone on both sides after execution is compared with the theoretical response data to obtain the execution residual data. The subsequent control actions and dynamic boundary data are then corrected based on the execution residual data.

2. The method for controlling the local heat equalization temperature field of the tank weld according to claim 1, characterized in that, The circumferential seam of the tank is divided into weld control sections according to the circumferential arc length or the longitudinal seam of the tank is divided according to the axial length, including: Read the tank radius, tank plate thickness, tank axial length, and tank support position from the tank geometry data; Read the weld type, weld centerline position, weld start position, weld end position, and welding direction from the weld path data; When the weld type is circumferential weld, the weld centerline is divided into multiple weld control segments according to the circumferential arc length, and the circumferential start point, circumferential end point, segment center position, segment number, preceding adjacent segment number and subsequent adjacent segment number of each weld control segment are recorded. When the weld type is longitudinal, the weld centerline is divided into multiple weld control segments according to the axial length, and the axial start point, axial end point, segment center position, segment number, preceding adjacent segment number and subsequent adjacent segment number of each weld control segment are recorded.

3. The method for controlling the local uniform heating temperature field of the tank weld according to claim 1, characterized in that, Write the temperature measurement point number, unfolding position, and weld centerline side into the temperature measurement mapping table, including: Read the temperature measurement point number, temperature measurement location, temperature measurement direction, temperature measurement coverage area, and time synchronization relationship from the temperature measurement and positioning data; By projecting the temperature measurement position onto the start and end range of the weld control section, the weld control section to which the temperature measurement point belongs and the development position are obtained. Using the weld centerline as the lateral reference, one side of the welding direction is fixed as the first side, and the opposite side is fixed as the second side. Temperature measurement points located within a preset width range on both sides of the weld centerline are recorded as the area near the centerline. Write the temperature measurement point number, the corresponding weld control section, the unfolding position, the weld centerline side, and the effective coverage period into the temperature measurement mapping table.

4. The method for controlling the local heat equalization temperature field of the tank weld according to claim 1, characterized in that, The temperature measurement values ​​of the current sampling period are converted into unfolded temperature points with segment number, side, unfolded position, and confidence marker according to the temperature measurement mapping table, including: The temperature measurement point number of the current sampling period is used to read the corresponding weld control section, weld centerline side and unfolding position from the temperature measurement mapping table; The temperature value of the current sampling cycle, the sampling time, the control section of the weld, the side of the weld centerline, and the development position are combined to form the development temperature point; Within the sliding sampling window, when the temperature measurement value is within the temperature measurement range, there are valid temperature measurement values ​​in both the previous and next sampling cycles, and the unfolding position is located within the effective area of ​​the corresponding weld control section, the unfolding temperature point is marked as the first reliable temperature point. When the measured temperature value is within the temperature measurement range but is adjacent to a missing sampling point or a temperature point outside the boundary, the expanded temperature point will be marked as the second reliable temperature point. When a temperature measurement value is missing, the temperature measurement point number cannot be matched in the temperature measurement mapping table, or the temperature measurement value exceeds the temperature measurement range, the expanded temperature point will be marked as an invalid temperature point.

5. The method for controlling the local uniform heating temperature field of the tank weld according to claim 1, characterized in that, Within the same weld control section, temperature points on the first and second sides are read along the direction away from the weld centerline to obtain the width of the double-sided heat-affected zone, including: Within the same weld control section, the development temperature points are divided into the first side temperature point, the second side temperature point, and the temperature point near the center line according to the side of the weld center line. Arrange the temperature points on the first and second sides in a direction away from the weld centerline, and delete invalid temperature points. When the temperature point on the outer side of the same side is lower than the lower limit of the heat-affected zone edge temperature record in the historical qualified weld, and the adjacent temperature point on the inner side is within or higher than the lower limit of the heat-affected zone edge temperature record in the historical qualified weld, the position of the adjacent temperature point on the inner side is recorded as the boundary point of the heat-affected zone on the corresponding side. The width of the first heat-affected zone is obtained from the distance from the boundary point of the first heat-affected zone to the center line of the weld. The width of the second heat-affected zone is obtained from the distance from the boundary point of the second heat-affected zone to the center line of the weld. The difference between the widths of the two heat-affected zones is obtained from the widths of the first and second heat-affected zones.

6. The method for controlling the local uniform heating temperature field of the tank weld according to claim 1, characterized in that, The dynamic boundary data is obtained in the following way: Multiple candidate historical windows are obtained from the historical set of temperature points, the width of the historical bilateral thermally affected zone, and historical execution records; Count the number of missing temperature points, invalid temperature points, number of boundary direction flips, and execution failure flags for each candidate historical window; Candidate historical windows with no missing temperature points exceeding the integrity limit, no invalid temperature points exceeding the invalidity limit, no boundary direction flipping exceeding the flipping limit, and no execution failure flags are considered as valid stable windows. The dynamic boundary data of the first side is obtained from the width distribution range of the first side heat-affected zone in the effective stable window, and the dynamic boundary data of the second side is obtained from the width distribution range of the second side heat-affected zone in the effective stable window. The dynamic boundary data is used for comparison with the current width of the heat-affected zone and is not used in the initial identification of the boundary points of the current heat-affected zone.

7. The method for controlling the local uniform heating temperature field of the tank weld according to claim 1, characterized in that, The widths of the heat-affected zones on both sides of the current weld seam control section are compared with the widths on the same side and the corresponding dynamic boundary data of the preceding adjacent weld seam control sections to obtain boundary deviation data and section-level back-visit data, including: By comparing the width of the heat-affected zone on the first side of the current weld control section with the dynamic boundary data on the first side, the deviation direction and offset distance on the first side are obtained. By comparing the width of the heat-affected zone on the second side of the current weld control section with the dynamic boundary data on the second side, the deviation direction and offset distance on the second side are obtained. When the width of the heat-affected zone is greater than the outer boundary of the corresponding dynamic boundary data, the deviation direction is recorded as outward expansion; when the width of the heat-affected zone is less than the inner boundary of the corresponding dynamic boundary data, the deviation direction is recorded as insufficient; when the width of the heat-affected zone is within the corresponding dynamic boundary data, the deviation direction is recorded as normal. Compare the first side deviation direction of the current weld control segment with the first side deviation direction of the previous adjacent weld control segment to obtain the first side boundary back visit data. Compare the second side deviation direction of the current weld control segment with the second side deviation direction of the previous adjacent weld control segment to obtain the second side boundary back visit data. When the first-side deviation direction of the current weld control segment and the preceding adjacent weld control segment is both outward expansion, the segment-level back-visit data is recorded as the continuation of the first-side heat accumulation; when the second-side deviation direction of the current weld control segment and the preceding adjacent weld control segment is both outward expansion, the segment-level back-visit data is recorded as the continuation of the second-side heat accumulation; when both-side deviation directions are outward expansion, the segment-level back-visit data is recorded as the overall heat accumulation continuation; when the current weld control segment has outward expansion or insufficiency and the same-side deviation direction of the preceding adjacent weld control segment is normal, the segment-level back-visit data is recorded as the local offset of the current segment.

8. The method for controlling the local heat equalization temperature field of the tank weld according to claim 1, characterized in that, A heat-soaking status code is generated from boundary deviation data and segment-level backtracking data. Candidate process parameters with opposite directions are filtered out according to the heat input direction defined by the heat-soaking status code. The target process parameters are then obtained by combining the boundary deviation data of subsequent adjacent weld control segments, including: Boundary deviation direction code, bilateral equalization code and return persistence code are obtained from boundary deviation data, and hot history return code is obtained from segment-level return data. The homogenization status code is formed by combining the boundary deviation direction code, thermal history return code, bilateral equalization code, and return persistence code in that order. When the heat soaking status code corresponds to overall expansion or continuous expansion, candidate process parameters that increase welding current, increase welding voltage, decrease welding speed, increase wire feeding speed, or extend the heat source action time will be deleted. When the overall heat dissipation status code is insufficient, candidate process parameters such as reducing welding current, reducing welding voltage, increasing welding speed, reducing wire feed speed, or shortening the heat source action time will be deleted. When the heat spread status code corresponds to unilateral expansion, the candidate process parameters that would increase the effect of the heat source on the expansion side will be deleted. When the heat input direction corresponding to the candidate process parameters after screening is opposite to the deviation convergence direction of the subsequent adjacent weld control segment, the adjustment range of the candidate process parameters is reduced or they are limited to be executed only in the current weld control segment to obtain the target process parameters.

9. The method for controlling the local uniform heating temperature field of the tank weld according to claim 1, characterized in that, After executing the target process parameters, the width of the heat-affected zone on both sides after execution is compared with the theoretical response data to obtain the execution residual data. This execution residual data is then used to correct subsequent control actions and dynamic boundary data, including: The theoretical response data is obtained from the pre-execution boundary deviation data, target process parameters, and historical valid execution records. The theoretical response data includes the boundary change direction, response waiting period, and expected return segment. After the response waiting period, the width of the bilateral heat-affected zone after execution is reacquired and compared with the theoretical response data. When the width of the bilateral heat-affected zone moves along the boundary change direction in the theoretical response data and enters the expected return segment after execution, the execution residual data is recorded as convergence in the same direction, and subsequent control actions maintain the current direction and reduce the adjustment amplitude. When the width of the double-sided heat-affected zone moves along the boundary change direction in the theoretical response data after execution but does not enter the expected return section, and the target process parameters have been correctly read back, the execution residual data is recorded as insufficient response, and subsequent control actions maintain the current direction and increase the adjustment amplitude; When the width of the double-sided heat-affected zone moves along the boundary change direction in the theoretical response data after execution but does not enter the expected return section, and the target process parameters are not correctly read back, the same type of control action is suspended and the execution link verification is initiated. When the width of the double-sided heat-affected zone moves in the opposite direction to the boundary change direction in the theoretical response data after execution, the execution residual data is recorded as reverse divergence, the same-direction control action is canceled and the candidate process parameters are regenerated. When the residual data increases in the same direction over multiple consecutive control cycles, and the post-weld quality confirmation result of the subsequent effective stable window is qualified, the corresponding dynamic boundary data is moved outward or inward.

10. A system for controlling the local heat equalization temperature field of tank welds, characterized in that, include: The weld seam control section generation module is used to divide the tank circumferential seam into multiple weld seam control sections according to the circumferential arc length or the tank longitudinal seam according to the axial length. The temperature measurement mapping module is used to write the temperature measurement point number, unfolding position and weld centerline side into the temperature measurement mapping table; The expanded temperature point generation module is used to convert the temperature measurement value of the current sampling period into an expanded temperature point with segment number, side, expanded position and confidence mark according to the temperature measurement mapping table; The double-sided heat-affected zone boundary identification module is used to read the temperature points on the first and second sides along the direction away from the weld centerline within the same weld control section to obtain the width of the double-sided heat-affected zone. The segment-level comparison module is used to compare the width of the heat-affected zone on both sides of the current weld control segment with the width on the same side and the corresponding dynamic boundary data of the adjacent weld control segment in the previous sequence, so as to obtain boundary deviation data and segment-level back visit data. The homogenization status code generation module is used to generate homogenization status codes from boundary deviation data and segment-level back-visit data. The target process parameter generation module is used to filter out candidate process parameters with opposite directions according to the heat input direction defined by the heat soaking status code, and obtain the target process parameters by combining the boundary deviation data of the subsequent adjacent weld control sections. The control execution module is used to execute the target process parameters; The execution residual generation module is used to compare the width of the double-sided heat-affected zone after execution with the theoretical response data to obtain the execution residual data. The correction module is used to correct subsequent control actions and dynamic boundary data based on the execution residual data; The historical time-series database is used to store weld control sections, temperature measurement mapping tables, development temperature points, width of the double-sided heat-affected zone, boundary deviation data, section-level revisit data, heat soaking status codes, target process parameters, execution residual data, and dynamic boundary data.

Images