Search results

Flexible automated assembly is an emerging need in several industries. The purpose of this paper is to address the introduction of an innovative concept in flexible assembly: the fully flexible assembly system (F‐FAS).

After an analysis of the state of the art, the authors describe the proposed F‐FAS, from a layout, constitutional elements, functioning principles and working cycle point of view. Second, the authors compare the traditional FAS and the manual assembly system versus the proposed F‐FAS according to their throughput and unit production costs, deriving a convenience map as a function of the number of components used in assembly and of the efficiency of the F‐FAS. Finally, using a prototype work cell developed at the Robotics Laboratory of University of Padua, the authors validate the F‐FAS concept.

Results of the research indicate that the concept of full‐flexibility can be exploited to bring automation to a domain where traditional FAS are not competitive versus manual assembly. In fact, the F‐FAS outperforms both traditional FAS and manual assembly, in terms of unit direct production costs, when the size of the batch is small, the number of components used in assembly is large and the efficiency of the F‐FAS is reasonably high. The F‐FAS prototype demonstrated the possibility of working, for certain conditions (models/components/production mix), in the F‐FAS convenience area, highlighting the achievable cost reduction versus traditional assembly systems.

The novelty of the study lies in the F‐FAS concept, its performances in terms of flexibility, compactness, throughput and unit direct production costs. A prototype work cell validated the concept and demonstrated its viability versus traditional assembly systems, thanks to convenience analysis.

The capability of an artificial neural network to determine part pose by processing image data from the silhouette of a back‐lit part has been established in recently reported work. The chief benefit of this new approach is simplicity of training, which is important for flexible automated parts feeders. The objective of the work presented herein is to develop an effective and efficient method for determining the position and orientation of the parts to be used in training the neural network. Candidate methods were used to create sets of training data containing different numbers of images taken of each part in different patterns of position and orientation. For each set of training data, the neural network was trained and its pose recognition performance was empirically evaluated. Based on these empirical results, a method for generating training data is reported that ensures accurate performance of the trained neural network while requiring only a minimum amount of training data.

The development of a generic flexible assembly system involves the design, selection and integration of a number of different mechanical systems in order to develop an assembly system, which is capable of assembling a wide variety of products having an unknown specification. A specific system configuration being dependent on a variety of factors such as, product size, weight, component insertion direction, and manipulator geometry. This paper examines each of the factors that should be considered when designing a generic flexible assembly system and presents a novel generic flexible assembly system design.

This paper aims to provide a framework for the choice, design, set-up and management of a fully flexible assembly system (F-FAS). Many industrial applications for small batch productions require highly flexible automated manufacturing systems. Moreover, some extensions of the F-FAS concept are provided.

The paper reviews recent findings regarding the F-FAS with a top-down approach, and defines an integrated implementation framework. This framework is structured into three strictly correlated phases, and the presented procedure is organized to be readily used for new industrial applications. Practical applications are presented to show how the system can satisfy flexibility demands in a variety of cases.

The proposed framework is organized in three steps: convenience analysis of the F-FAS compared to a traditional flexible assembly system; an optimal design of the feeder; a choice of the set-up and sequencing algorithm yielding the highest throughput. Following these steps, the F-FAS can become an effective solution for small batch productions with frequent reconfigurations. However, due to the limited throughput, the system is not well suited for large batches.

The presented framework allows to implement an F-FAS for a given industrial application, and to evaluate its efficacy with respect to other assembly technologies. Moreover, with the same implementation framework, the F-FAS concept can be applied to production fields that are different from assembly, as shown by the provided examples. This represents an important element of originality and of interest for its strong practical implications in different production environments.

Flexible feeding is an emerging alternative to traditional part feeding methods. This alternative greatly enhances the versatility of a manufacturing workcell by using a robot manipulator and sophisticated sensing devices such as machine vision, thereby significantly reducing both cost and set up time. This article explores the benefits of a new model in PC‐based robot control, which makes the development of flexible feeders and similar applications much easier than using traditional robot programming environments. It also explores how a programming paradigm based on a well‐defined model of the workcell greatly simplifies both the logic of the application and the calibration of the physical machine.

Renault Machines Outils has been pulling together various expertise available within the Renault organisation in order to develop the flexible manufacturing concept (FMS) both for use within the company and for its outside customers.

This paper aims to focus on the novel design and development of an automatic feeding system which is capable of feeding cylindrical parts which are fragile and powdery in nature and possess asymmetrical features such as a groove near to one end.

It is an active feeder, performing its task without having to reject any feeding part by performing active orientation of feeding parts that are in the undesired orientation. This design incorporating active orientating capability is aimed at 100 percent feeding efficiency. The system is controlled and driven by a programmable logic controller and electropneumatics.

System evaluation results showed that the average jam rate is below 5 percent and the percentage of correctly orientated parts is above 95 percent. With enhancement and fine tuning, the system could become a very useful feeder for industry in the future.

The scope of this paper focuses on presentation of the design concept, development and evaluation of the feeder only and design calculations are not included.

This paper is of value to those who are involved in the manufacturing of small delicate and powdery engineering parts such as those providing performs to the semiconductor industry for encapsulation of integrated circuit chips.

Motivated by the problems of flexible manufacturing and parts feeding in sensorless assembly, a flat, level vibrating table is proposed as a means of orienting small parts for assembly. Ideas inspired by academic literature regarding juggling and dynamic manipulation are used to design an oscillatory motion of the table for simple 1 and 3 degree of freedom parts. Passive juggling, in which a part is repeatedly caught and tossed by the table without sensing or feedback, forms the fundamental manipulation strategy. The oscillation is intended to ensure attraction to a predetermined orientation from any initial orientation within a few seconds, without the need for sensing or fixtures. Preliminary simulation results suggest the feasibility of this approach.

The purpose of this paper is to discuss a linear vibratory part feeder for handling brake liners, typical sector-shaped components. Part feeders have been used in the industries for a long time to present the parts in a desired orientation. Berretty et al. (1999) discussed a class of mechanical filters that are capable of removing polygonal sections from the track of the feeder which are referred to as traps. The traps eliminate or reorient the parts until they reach the final desired orientation. A part feeder was developed using traps, to reorient the sector-shaped part to desired orientation. The desired orientation was the most probable natural resting orientation. The trap was mounted on a linear vibratory feeder. The adaptive part feeder developed was capable of identifying the size of the incoming part and adjust the trap to accommodate that. This set-up eliminates the use of different traps for different-sized sector-shaped parts and wastage of productive time in changing the traps for different sizes. A regression model was developed to predict the conveying velocity of part on the feeder.

A part feeder was developed using traps, to reorient the sector-shaped part to desired orientation. Acrylic material was found to be suitable for trap compared to aluminium. The adaptive part feeder developed was capable of identifying the size of the incoming part using proximity sensors. Depending on the size of the incoming part, the track width was adjusted dynamically with the help of a stepper motor, rack and pinion arrangement. A regression model was developed to predict the conveying velocity.

Typical brake liners in the size range of 40-60 mm (radius) were considered for developing the adaptive part feeder. Based on performance studies, the acrylic trap was found better than aluminium traps. The appropriate frequency and amplitude of vibration for maximum conveying velocity of the adaptive part feeder were found experimentally. Regression equation was developed to determine the conveying velocity based on input frequency and amplitude. The regression results were found to be in close agreement with the experimental results.

The developed part feeder is suitable for handling sector-shaped parts only.

This paper demonstrates an inexpensive adaptive part feeding device for handling sector-shaped parts which can be extended for handling other asymmetric parts also.

We study a programmable robotic part feeder that relies on a sequence of three conveyor belts to singulate and re‐circulate parts. In industrial practice, belt speeds are set in an ad hoc fashion. Experience with real feeders reveals that throughput can suffer owing to: starvation where no parts are visible to the camera; and saturation, where too many parts are visible, which prevents identifying part pose or grasping due to obstruction by nearby parts. This motivates our search for a systematic approach to setting belt speeds. Our goal is to optimize throughput, measured in terms of how many parts per second are delivered from the robotic feeder. We describe a 1D model of the belts with a Poisson arrival process to stochastically model how belt speeds affect throughput. Initially, we study the finite case where N parts are placed into the feeder and re‐circulated until they are all delivered by the robot. Our first insight is that the vision belt should be run at maximum achievable velocity. We run simulations to empirically determine optimal buffer belt velocity as a function of lot size. Finally, we develop a theoretical model for the case where N = ∞ which approximates common usage where the buffer is replenished before it becomes empty. From this model, we derive the optimal buffer belt velocity and show that it produces throughput five times greater than that achieved with ad hoc settings.