Optimization of a clamping concept based on machine learning

Feng, Qi; Maier, Walther; Stehle, Thomas; Möhring, Hans‐Christian

doi:10.1007/s11740-021-01073-z

Cited by 15 publications

(7 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the successes of our work, there are some limitations that need to be addressed. The first limitation, one which is prevalent across all fixture planning literature, is the verification of the plans outside of simulation on hardware [2], [12], [14], [16], [26]. The main reason for this limitation is two-fold: firstly, acquiring the components mentioned in both this work and the work cited prior and developing an experimental test-bed remains costly in an academic setting; secondly, measuring the deformations exhibited by all the fixture planners requires high resolution sensors.…”

Section: Discussionmentioning

confidence: 99%

“…Recent methods in fixture design planning have utilised the data-based approach of machine learning techniques to find optimal fixture plans. Early work sought to mimic the methodology of CBR and RBR by using pattern recognition through neural networks to find suitable similar fixtures [15], with further work looking at using other supervised methods of machine learning [16]. However, supervised methods that rely on prior data collection fall victim to similar drawbacks as CBR and RBR due to the lack of guarantee of optimal fixturing plans.…”

Section: A Fixture Layout Planningmentioning

confidence: 99%

See 1 more Smart Citation

Decision Making For Multi-Robot Fixture Planning Using Multi-Agent Reinforcement Learning

Canzini,

Noguera,

Pope

et al. 2024

Preprint

View full text Add to dashboard Cite

Within the realm of flexible manufacturing, fixture layout planning allows manufacturers to rapidly deploy optimal fixturing plans that can reduce surface deformation that leads to crack propagation in components during manufacturing tasks. The role of fixture layout planning has evolved from being performed by experienced engineers to computational methods due to the number of possible configurations for components. Current optimisation methods commonly fall into sub-optimal positions due to the existence of local optima, with data-driven machine learning techniques relying on costly to collect labelled training data. In this paper, we present a framework for multi-agent reinforcement learning with team decision theory to find optimal fixturing plans for manufacturing tasks. We demonstrate our approach on two representative aerospace components with complex geometries across a set of drilling tasks, illustrating the capabilities of our method; we will compare this against state of the art methods to showcase our method’s improvement at finding optimal fixturing plans with 3 times the improvement in deformation control within tolerance bounds.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: A Fixture Layout Planningmentioning

confidence: 99%

Decision Making For Multi-Robot Fixture Planning Using Multi-Agent Reinforcement Learning

Canzini,

Noguera,

Pope

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…The effectiveness of the proposed method has been validated through practical positioning layouts. [ 21 ] Feng et al [ 22 ] proposed a new approach to optimize clamping concepts and fixture design using machine learning instead of finite element methods, which aims to reduce manufacturing errors and achieve higher fixture stiffness and machining accuracy. Yang et al [ 23 ] adopted a method based on Kriging surrogate models and multi‐objective bat algorithm to address the issues of excessive computational cost and multi‐objective attributes in finite element analysis during the optimization process of sheet metal fixture positioning layout.…”

Section: Introductionmentioning

confidence: 99%

An Improved Simulated Annealing Algorithm in Dataset Domain for Optimizing Robust Workpiece Fixture Layout

Ding

Dong

Gao

et al. 2023

Advcd Theory and Sims

View full text Add to dashboard Cite

The robust layout of the fixture on the workpiece solves the problem of workpiece fixing well. Aiming at the problem of workpiece position error caused by fixture source error on workpiece surface, an improved simulated annealing algorithm based on dataset domain is adopted to reduce the workpiece position error caused by fixture locator deviation as much as possible. To find a better robust layout way, a point differentiation method is proposed to discretize continuous workpieces to obtain the dataset domain. Meanwhile, two objective functions considering curvature and non‐curvature are used to evaluate the placement scheme of locators. A point differentiation method is proposed to discretize continuous workpieces to obtain the dataset domains. Meanwhile, two objective functions considering curvature and non‐curvature are used to evaluate locator layout schemes. To improve the performance of the algorithm, an AHP‐based multi‐attribute decision‐making method is used to find a set of optimal parameters to configure the algorithm. The benchmark example and practical verification show that the locator obtained by the algorithm proposed is less discrete and denser, which solves the problem of fixture positioning very well, and the position error of the workpiece is effectively reduced.

show abstract

“…Li et al [29] used whale optimization algorithm to reduce the clamping deformation of the curved thin-walled parts. Feng et al [30] proposed the recent machine learning-based prediction model of the workpiece deformation calculator using extreme gradient boosting (XGBoost) method.…”

Section: Introductionmentioning

confidence: 99%

“…Then, optimizer should reach the optimal results using the limited number of trials. On the other hand, the second group of studies are based on the combination of machine learning developed model and optimization technique [15,17,18,28,30]. Using these developed methods based on machine learning technique, the recalculation of the workpiece deformation can be achieved in couple of seconds.…”

Section: Introductionmentioning

confidence: 99%

Optimization of fixture locating layout design using comprehensive optimized machine learning

Qazani

Parvaz

Pedrammehr

2022

Int J Adv Manuf Technol

View full text Add to dashboard Cite

Fixtures are commonly employed in production as work holding devices that keep the workpiece immobilized while machined. The workpiece’s deformation, which affects machining precision, is greatly influenced by the positioning of fixture elements around the workpiece. By positioning the locators and clamps appropriately, the workpiece’s deformation might be decreased. Therefore, it is required to model the fixture–workpiece system’s complicated behavioral relationship. In this study, long short-term memory (LSTM), multilayer perception (MLP), and adaptive neuro-fuzzy inference system (ANFIS) are three machine-learning approaches employed to model the connection between locator and clamp positions and maximum workpiece deformation throughout end milling. The hyperparameters of the developed ANFIS, MLP, and LSTM are chosen using the evolutionary algorithms, including genetic algorithm (GA), particle swarm optimization (PSO), butterfly optimization algorithm (BOA), grey wolf optimization (GWO), and wolf optimization algorithm (WOA). Among developed methods, MLP optimized using BOA (BOA-MLP) reached the highest accuracy among all developed models in predicting the response surface. The developed model had a lower computational load than the final element model in calculating the response surface during the machining process. At the final step, the prementioned five evolutionary algorithms were implemented in the developed BOA-MLP to extract the optimal parameters of the fixture to decrease the deflection of the workpiece throughout the machining. The proposed method was modeled in MATLAB. The outcomes showed that the mentioned model was efficient enough compared with the previous method, such as optimized response surface methodology in the point view of 0.0441 μm lower workpiece deflection.

show abstract

Optimization of a clamping concept based on machine learning

Cited by 15 publications

References 58 publications

Decision Making For Multi-Robot Fixture Planning Using Multi-Agent Reinforcement Learning

Decision Making For Multi-Robot Fixture Planning Using Multi-Agent Reinforcement Learning

An Improved Simulated Annealing Algorithm in Dataset Domain for Optimizing Robust Workpiece Fixture Layout

Optimization of fixture locating layout design using comprehensive optimized machine learning

Contact Info

Product

Resources

About