Rapid sequence optimization of spot welds for improved geometrical quality using a novel stepwise algorithm

To reduce welding deformation during the automated welding of thick-walled pipes in shipbuilding and thereby improve welding quality, a segmented multi-layer multi-pass welding sequence optimization and process improvement strategy is proposed. Firstly, based on a welding model for thick-walled pipes, a multi-layer multi-pass welding trajectory equation is established. A double-ellipsoidal moving heat source is adopted to design a circular multi-layer multi-pass double-ellipsoidal heat source model. Secondly, three circular pipe workpieces with different wall thicknesses are selected, and four segmented welding sequences are simulated using welding finite element analysis (FEA). Finally, based on the optimal segmented welding sequence, the welding process is improved, and optimal welding process parameters are determined based on deformation and residual stress analysis. The results of the segmented multi-layer multi-pass welding sequence optimization show that the skip-symmetric welding method yields the best results for thick-walled circular pipes. Compared to other welding sequences, it reduces welding deformation by an average of 6.50% and welding stress by an average of 5.37%. In addition, process improvement tests under the optimal welding sequence indicate that the best welding quality is achieved under the following conditions: for 10 mm thick pipes—200 A current, 24 V voltage, and 11.5 mm/s welding speed; for 15 mm thick pipes—215 A, 24.6 V, and 10 mm/s; and for 20 mm thick pipes—225 A, 25 V, and 11 mm/s.

Disassembly planning and sequencing play an important role in recycling a fast-growing number of end-of-life products. Optimal sequences can effectively reduce carbon emissions and save natural resources in the remanufacturing industry. Considering the development of intelligent manufacturing technology, this work deals with the optimization problem of selective disassembly sequences with an objective of maximizing disassembly profit. Disassembly sequences are generated based on AND/OR graphs. After setting up an environment matrix based on such graphs, this proposes a Q-learning technique to find an selective optimal disassembly sequence. The algorithm is applied to real-life disassembly cases. Experimental results show that the algorithm is superior a popularly-used genetic algorithm (GA) in both computing speed and solution quality through their various comparisons.

Incorporating domain knowledge into reinforcement learning to expedite welding sequence optimization

Code space is a critical issue facing designers of software for embedded systems. Many traditional compiler optimizations are designed to reduce the execution time of compiled code, but not necessarily the size of the compiled code. Further, different results can be achieved by running some optimizations more than once and changing the order in which optimizations are applied. Register allocation only complicates matters, as the interactions between different optimizations can cause more spill code to be generated. The compiler for embedded systems, then, must take care to use the best sequence of optimizations to minimize code space.Since much of the code for embedded systems is compiled once and then burned into ROM, the software designer will often tolerate much longer compile times in the hope of reducing the size of the compiled code. We take advantage of this by using a genetic algorithm to find optimization sequences that generate small object codes. The solutions generated by this algorithm are compared to solutions found using a fixed optimization sequence and solutions found by testing random optimization sequences. Based on the results found by the genetic algorithm, a new fixed sequence is developed to reduce code size. Finally, we explore the idea of using different optimization sequences for different modules and functions of the same program.

Geometric variation is one of the sources of quality issues in a product. Spot welding is an operation that impacts the final geometric variation of a sheet metal assembly considerably. Evaluating the outcome of the assembly, considering the existing geometrical variation between the components can be achieved using the Method of Influence Coefficients (MIC), based on the Finite Element Method (FEM). The sequence, with which the spot welding operation is performed, influences the final geometric deformations of the assembly. Finding the optimal sequence that results in the minimum geometric deformation is a combinatorial problem that is experimentally and computationally expensive. For an assembly with N number of welds, there are N! possible sequences to perform the spot welding operation. Traditionally, spot welding optimization strategies have been to simulate the geometric variation of the spot-welded assembly after the assembly has been positioned in an inspection fixture, using an appropriate measure of variation. In this approach, the calculation of deformation after springback is one of the most time-consuming steps. In this paper, the cause of variation in the deformations after the springback, between different sequences is identified. The relative displacements of the weld points in the assembly fixture, when welded in a sequence, is the source of such behavior. Capturing these displacements leads to large time savings during sequence optimization. Moreover, this approach is independent of the inspection fixture. The relative weld displacements have been evaluated on two sheet metal assemblies. The sequence optimization problem has been solved for the two assemblies using this approach. The optimal sequence, the corresponding final assembly deformations, and the time-consumption have been compared to the traditional approach. The results show a significant correlation between the weld relative displacements in the assembly fixture, and the assembly deformation in the inspection fixture. Considering the relative weld displacement makes each assembly evaluation less time-consuming, and thereby, sequence optimization time can be reduced up to 30%, compared to the traditional approach.

Geometrical variation is one of the sources of quality issues in a product. Spot welding is an operation that impacts the final geometrical variation of a sheet metal assembly considerably. Evaluating the outcome of the assembly, considering the existing geometrical variation between the components, can be achieved using the method of influence coefficients (MICs), based on the finite element method (FEM). The sequence with which the spot welding operation is performed influences the final geometrical deformations of the assembly. Finding the optimal sequence that results in the minimum geometrical deformation is a combinatorial problem that is experimentally and computationally expensive. Traditionally, spot welding sequence optimization strategies have been to simulate the geometrical variation of the spot-welded assembly after the assembly has been positioned in an inspection fixture. In this approach, the calculation of deformation after springback is one of the most time-consuming steps. In this paper, a method is proposed where the springback calculation in the inspection fixture is bypassed during the sequence evaluation. The results show a significant correlation between the proposed method of weld relative displacements evaluation in the assembly fixture and the assembly deformation in the inspection fixture. Evaluating the relative weld displacement makes each assembly simulation less time-consuming, and thereby, sequence optimization time can be reduced by up to 30%, compared to the traditional approach.

The representation of the disassembly sequence is a key issue in maintenance planning. It involves a highly constrained combinatory problem, which is coupled with varying start and end nodes. These start and end nodes are dependent upon the nature of a maintenance task. In this paper, different representation and modelling schemes for disassembly sequence planning are first reviewed. Then, a novel representation scheme for disassembly sequence, which is generic and can be used to represent both the geometrical and precedence constraints dynamically in product disassembly, is proposed. Based on such a representation scheme, the process for the determination of possible disassembly sequences can be simplifled. By taking into consideration disassembly constraints, the optimal disassembly sequence in relation to the component to be maintained (target component) can be quickly derived. This is achieved by pruning the search space of disassembly sequences, grouping related components into subassemblies, and identifying free components to facilitate disassembly operations. Subsequently, the optimal disassembly sequence in relation to the target component can be obtained using genetic algorithms. In this manner, the disassembly sequences for a complex product comprising a relatively large number of components can be derived within a short time. A case study is used to illustrate the effectiveness of the representation scheme. Comparisons are made using the same case study with the AND/OR graph representation and the Petri net approach for disassembly sequence planning. The results show that the proposed representation scheme is simpler and is more efficient than the rest.

Stiffened plate panel is the major structural part of a fabrication industry where fillet welding joint is one of the most important fabrication techniques. Large stiffened structures are generally joined by several welding passes which generates thermal stresses and angular deformation. Tensile residual stresses which are generated due to welding in the weld region may lead to early failure of the structure when subjected to cyclic loading. The weld-induced residual distortion causes dimensional inaccuracy and needs rework to achieve the desired shape. Use of multiple welding passes without any optimized welding sequences typically leads to an increased degree of nonuniform heating and cooling, i.e., creating complex welding residual stress and angular deformation in the structure. In this present study, the effect of four different welding sequences on submerged arc welded fillet joint has been studied. A finite element-based numerical model has been developed to predict the thermal profile, welding residual stress, and angular deformation. The developed model considers temperature-dependent material property and material deposition by using element death and birth technique. The results have been compared with experimental one. In the effect of welding sequence on residual stress, angular deformation has been studied. Thus, the developed model presents the effect of welding sequence on the weld induced residual stresses and distortions which provide one of the most optimal welding sequence for enhanced fabrication process.

Based on the local-global mapping model of the side-beam of low-pressure cylinder frame, the welding sequence of Q235-C steel was studied by finite element method, and the welding residual deformation under different welding schemes was obtained. By comparison, it can be seen that the option 2 from the Circumferential-radial welding is superior to option 1 of the circumferential-radial simultaneous welding. It is also superior to option 3 and option 4, which are radial-circumferential welding and radial-circumferential-radial welding, respectively. In addition, the whole process predicts and controls the welding deformation by numerical simulation, which provides a numerical basis for reducing and controlling the welding deformation of the side beam. The optimal welding sequence with the minimum welding deformation is given during the actual welding process.

8 - Mitigating welding residual stress and distortion

Delta Robots are used in industry for light weight material handling and sorting. This paper presents a sequence optimizing methodology for a color and size sorting delta robot. It finds the optimum path in the task space to perform an industry emulated scenario. An OpenCV-Python program was developed to sort objects according to their colors and sizes. The static positional coordinates of the objects in the robot workspace are obtained using the program. Genetic algorithm is used for pick-and-place sequence optimization to ensure that the sorting process is performed in the shortest possible path. The static positional coordinates are used to calculate the fitness. Single point crossover and mutation are applied with elitist selection when the current generation evolves to the next generation. The genetic algorithm ensures that the sequence of pick-and-place converges to the highest fitness in minimum number of generations reducing the computational time.

Energy savings is a major challenge in distillation operations. However, there is still one problem, which is how do we improve the energy efficiency of the existing distillation column systems without major modifications. Recently, a new energy efficient distillation columns methodology that will be able to improve energy efficiency of the existing separation systems without having major modifications has been developed. Therefore, the objective of this paper is to present a new improvement of the existing methodology by designing an optimal sequence of energy efficient distillation columns using a driving force method. Accordingly, the methodology is divided into four hierarchical sequential stages: i) existing sequence energy analysis, ii) optimal sequence determination, iii) optimal sequence energy analysis, and iv) energy comparison and economic analysis. The capability of this methodology is tested in designing an optimal synthesis of energy efficient distillation columns sequence of an aromatics separation unit. The existing aromatics separation unit consists of six compounds (Methylcyclopentane (MCP), Benzene, Methylcyclohexane (MCH), Toluene, m-Xylene and o-Xylene) with five direct sequence distillation columns being simulated using a simple and reliable short-cut method and rigorously tested within an Aspen HYSYS® simulation environment. The energy and economic analyses show that the optimal sequence determined by the driving force method has a better energy reduction with a total of 6.78% energy savings and a return of investment of 3.10 with a payback period of 4 months. It can be concluded that, the sequence determined by the driving force method is not only capable in reducing energy consumption, but also has a better economic cost for an aromatic separation unit

A hybrid algorithm based on MOSFLA and GA for multi-UAVs plant protection task assignment and sequencing optimization

On May 22, 2021, the Expert Board met in St. Petersburg to discuss their position on immunological aspects of determination of an adequate biological treatment sequence for inflammatory bowel diseases (IBD). The Expert Board aimed at discussion of current strategies, development of a consensus on determination of an adequate biological treatment sequence for IBD. The main topics of the agenda were the contribution of immune system to the pathophysiology of Crohn's disease, ulcerative colitis and their complications, efficacy of genetically engineered biological agents (GEBA) at various stages of IBD management. Participation of the leading Russian experts in IBD, as well as involvement of other specialties, made it possible to consider the topic by a multidisciplinary team, with an in-depth analysis of IBD pathophysiology, to better understand the course of the disease in some contradictory situation, for instance, when clinical remission is not associated with an endoscopically confirmed remission. One of the expected effects of this Expert Board meeting would be an improvement of GEBA administration in clinical practice, mostly due to the modification of clinical guidelines. This would ascertain and confirm the algorithms for GEBA administration for IBD, including the optimal treatment sequence depending on an agent’s mechanism of action and the patient profile. The clarification of the optimal GEBA sequence in the clinical guidelines could lead to more frequent GEBA administration in local medical clinics and institutions in the regions, where GEBA are used insufficiently due to little experience and absence of their precise positioning in the clinical guidelines.

A scatter search algorithm for stacking sequence optimisation of laminate composites