Automatic recognition of machining features using artificial neural networks

Sunil, V. B.; Pande, Shivam

doi:10.1007/s00170-008-1536-z

Cited by 76 publications

(33 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chakraborty and Basu (2006) used binary coding to recognise simple features. Sunil and Pande (2009) proposed 12-node vector coding to recognise a wide range of isolated complex features.…”

Section: Neural Network-based Feature Recognitionmentioning

confidence: 99%

A review of machining feature recognition methodologies

Verma

Rajotia

2010

International Journal of Computer Integrated Manufacturing

View full text Add to dashboard Cite

In a continuing quest to decrease the time interval between conceptualisation of a product and its first production, the use of information technology in design, analysis and manufacturing practice has been actively researched. The design engineer designs a part and sends the final design to the manufacturing engineer, who re-interprets the design and plans the manufacturing activities to produce the part. These two sections generally work in isolation from each other, resulting in high lead-time, duplication of data, inconsistent product data and sometimes redesign of a product. Feature recognition is a process of reinterpreting a design model database for automating downstream manufacturing activities. Active research in this field has developed numerous techniques such as syntactic pattern recognition, graph theory, volume decomposition, artificial intelligence and hint-based, and neural network-based systems. This paper presents a critical review of strengths and weaknesses of these approaches.

show abstract

“…Chakraborty and Basu (2006) used binary coding to recognise simple features. Sunil and Pande (2009) proposed 12-node vector coding to recognise a wide range of isolated complex features.…”

Section: Neural Network-based Feature Recognitionmentioning

confidence: 99%

A review of machining feature recognition methodologies

Verma

Rajotia

2010

International Journal of Computer Integrated Manufacturing

View full text Add to dashboard Cite

show abstract

“…Over the past decades, a lot of researches on this feature recognition technology have been performed at home and abroad, and great progress has been made. At present, the processing feature recognition methods are mainly divided into the following categories: The first type of method performs pattern matching based on the boundary conditions [3], such as based on the graph [4], based on the trace [5], based on the rule [6], based on the neural network [7], and so on. The second type of method is based on volume decomposition.…”

Section: Introductionmentioning

confidence: 99%

Recent Research and Prospect on Feature Recognition of Three-dimensional Model

Shi¹,

Deng²,

Zhong³

2018

Proceedings of the 2018 3rd International Conference on Control, Automation and Artificial Intelligence (CAAI 2018)

View full text Add to dashboard Cite

A feature recognition technique is the intelligent interface of CAD/CAPP. Through extracting and recognizing the tolerance and surface roughness, the design features such asholes, steps and other design features in a 3D model of parts, the CAD/CAPP can be combined efficiently, and then the CAD/CAPP/CAM system can be integrated. This technique also helps to shorten the cycle of product development, especially process preparation, and reduce the design cost. In this paper, the 3D model features recognition approaches about boundary pattern matching, volume composition, manufacturing resources, and hybridization method are reviewed, and compared in terms of identifying efficiency, identifying accuracy, identifying flexibility, machinability of features, and utilization of manufacturing resource information in a process environment is conducted on the 3D model feature recognition approaches. The main issues of machining feature recognition of the part model are also discussed, and the direction of future researches in this field is presented. It is pointed out that the development of an efficient, accurate, flexible, automatic and intelligent machining feature recognition system will become an important research topic in the field of computer integrated manufacturing in the future.

show abstract

“…Numerous techniques have been developed, such as graph-based feature recognition (Han, et al, 2000), volume decomposition-based feature recognition (Sakurai, 1995), rule-based feature recognition (Han, et al, 1998), hint-based feature recognition (Verma, et al, 2008), and neural network-based feature recognition (Sunil, et al, 2009). Among all these techniques, the graph-based approach is regarded as the most successful method, and it is firstly reported by Joshi and Chang (Joshi, et al, 1988).…”

Section: Introductionmentioning

confidence: 99%

Graph based automatic process planning system for multi-tasking machine

Zhu

Kato

Tanaka

et al. 2015

JAMDSM

View full text Add to dashboard Cite

Multi-tasking machine is capable of performing both milling and turning operations, it contributes to highly efficient machining and space conservation. However, prior to machining a lot of lead time is consumed in deciding efficient process plan, and generating machining tool path. Although the current CAM systems are highly integrated, the efficiency of the generated tool path is highly relied on the experience of the CAM programmer. In this research, an automatic process planning system for multi-tasking machine was developed. It is capable of recognizing manufacturing features and deciding efficient process plan from CAD model automatically. In this developed system, the CAD model is described as Attributed Adjacency Graph (AAG), and each feature is defined by AAG and its geometrical properties. Totally 8 milling features and 9 turning features can be recognized. The optimal machining plan is calculated based on machining cost evaluation. In addition, in this research a new method based on subfeature combination is proposed in order to recognize intersecting features. Furthermore, in this research the connection relationship of each feature is classified, and machining priority is assigned to adjacent features. It prevents the time consuming evaluation for checking all possible machining sequences. Finally, according to the experiment results, it is confirmed this developed system is capable of obtaining optimal machining plan properly and rapidly. IntroductionMulti-tasking machines have the capability to perform a variety of machining processes, such as turning, milling, drilling, et al., in a single setup. By using multi-tasking machine, it significantly reduces the setup time, and improves the machining accuracy and quality, because there is no need to move the workpiece across multiple machines. However, due to the complexity of multi-tasking machine, a lot of time is necessary for processing planning before actual machining. Usually, after a 3D model is designed in CAD (Computer Aided Design) system, CAM (Computer Aided Manufacturing) system is used to generate the NC (Numerical Control) data for machining. Although the current CAM systems are highly integrated, how to determine the machining feature, cutting tool, cutting condition and machining sequence is still a labor intensive task (Oba et al., 2010). For each single product, a long lead time is consumed to generate the cutting tool path, and the efficiency of the tool path is highly relied on the CAM programmer's experience. It leads to low productivity in high-mix and low-volume production. Therefore, a CAPP (Computer Aided Process Planning) system, which can automatically extract the manufacturing features from the CAD model and plan the machining process, has long been expected.Feature recognition and process planning are two of the most important components in CAPP system. Feature recognition is a process of reinterpreting a solid model as sets of manufacturing features, and process planning is a process to determine the optimal machin...

show abstract

Automatic recognition of machining features using artificial neural networks

Cited by 76 publications

References 26 publications

A review of machining feature recognition methodologies

A review of machining feature recognition methodologies

Recent Research and Prospect on Feature Recognition of Three-dimensional Model

Graph based automatic process planning system for multi-tasking machine

Contact Info

Product

Resources

About