Automated spray path planning based on Bayesian optimization and ant colony optimization

Automated spraying of forging release agents faces significant challenges, including inadequate lubricant coverage and excessive material waste due to complex mold geometries and dynamic production requirements. This study proposes a dynamic path-planning method based on ant colony optimization (ACO...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of advanced manufacturing technology Ročník 139; číslo 11-12; s. 5491 - 5509
Hlavní autoři: Peng, Ta-Jen, Chen, Peng-Jen, Su, Ting-Lun, Lin, Chung-Chieh
Médium: Journal Article
Jazyk:angličtina
Vydáno: London Springer London 01.08.2025
Springer Nature B.V
Témata:
Accuracy
Algorithms
Aluminum
Ant colony optimization
Assembly lines
Automation
Bayesian analysis
CAE) and Design
Computer-Aided Engineering (CAD
Configurations
Demand
Design
Ductility
Efficiency
Engineering
Evaporation rate
Forging
Gaussian process
High temperature
Industrial and Production Engineering
Lubricants & lubrication
Manufacturing
Mechanical Engineering
Media Management
Molds
Optimization
Original Article
Parameters
Path planning
Pheromones
Product quality
Release agents
Research methodology
Resource utilization
Robotics
Robots
Spraying
Temperature
Automated spraying
Intelligent manufacturing
Ant colony optimization
Forging
Bayesian optimization
Dynamic path planning
ISSN:0268-3768, 1433-3015
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Automated spraying of forging release agents faces significant challenges, including inadequate lubricant coverage and excessive material waste due to complex mold geometries and dynamic production requirements. This study proposes a dynamic path-planning method based on ant colony optimization (ACO) combined with Bayesian optimization (BO) to effectively address these challenges. The proposed approach dynamically optimizes critical algorithm parameters to effectively enhance spray coverage and reduce loss function values. Initially, the slope and height maps of the mold surface are generated using CAD software, quantifying the lubricant demand priorities of different regions. These data are then integrated into the pheromone update and path search processes of the ACO algorithm, significantly increasing pheromone accumulation in high-demand areas and enhancing spraying coverage efficiency. To overcome limitations of the ACO algorithm related to local optimality or prolonged trial-and-error parameter tuning, Bayesian optimization iteratively searches for optimal configurations of critical parameters (such as pheromone evaporation rate and heuristic weight). By continuously updating algorithm configurations based on loss function observations and Gaussian process modeling, the method ultimately converges to near-optimal solutions characterized by high coverage and effective resource utilization. Experimental results demonstrate that the optimized spraying system ensures lubricant stability under high-temperature forging conditions while substantially reducing redundant spraying in low-demand regions. By dynamically balancing exploration and exploitation, the loss function value decreased from an initial range of approximately 0.5–0.6 to below 0.06, confirming significant improvements in quality efficiency and material savings. The outcomes significantly enhance production yield, reduce material waste, and provide essential insights for advancing automation and smart manufacturing in the forging industry.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0268-3768
1433-3015
DOI:10.1007/s00170-025-16162-x