Kurapati Venkatesh

Design of artificial neural networks for tool wear monitoring

An on-line scheme for tool wear monitoring using artificial neural networks (ANNs) has been proposed. Cutting velocity, feed, cutting force and machining time are given as inputs to the ANN, and the flank wear is estimated using the ANN. Different ANN structures are designed and investigated to estimate the tool wear accurately. An existing analytical model is used to obtain the data for various cutting conditions in order to eliminate the huge cost and time associated with generation of training and evaluation data. Motivated by the fact that the tool wear at a given instance of time depends on the tool wear value at a previous instance of time, memory is included in the ANN. ANNs without memory, with one-phase memory, and with two-phase memory are investigated in this study. The effect of various training parameters, such as learning coefficient, momentum, temperature, and number of hidden neurons, on these architectures is studied. The findings and experience obtained should facilitate the design and implementation of reliable and economical real-time systems for tool wear monitoring and identification in intelligent manufacturing.

Object-oriented design of FMS control software based on object modeling technique diagrams and Petri nets

Abstract Control software is very important to realizing the full benefits of flexible manufacturing systems. This paper highlights the difficulties in developing such software and proposes an object-oriented design (OOD) method using object modeling technique (OMT) diagrams and Petri nets (PNs). OOD is used to design reusable and easily maintainable software. OMT diagrams are used to represent explicitly different kinds of static relations, such as generalization, aggregation, and association, among the objects in an FMS. PNs are used to model the dynamic behavior of the objects and the entire FMS and to conduct performance analysis. The use of PNs also leads to a method to identify the data structures and operations of software objects. The proposed method is illustrated through an FMS example. Its capability to support reusability, extendibility, and modifiability of the resulting control software design is also demonstrated when the FMS specifications change.

Augmented timed petri nets for modeling, simulation, and analysis of robotic systems with breakdowns

Abstract Flexible manufacturing and assembly systems consists of machines, robots, and automated guided vehicles aimed at meeting dynamically changing market needs. Numerous asynchronous concurrent actions involved in these systems make analysis of them difficult. Breakdowns of system components further complicate investigation of issues related to design, performance optimization, and control. This paper proposes a new class of modeling tools called augmented timed Petri nets (ATPNs) for modeling and analyzing robotic assembly systems with breakdowns. These models aid designers in better understanding the concurrency, synchronization, and sequential relations involved in breakdown handling and in system simulation for performance analysis. A flexible assembly system consisting of three robots with various breakdown rates is used to illustrate modeling, simulation, and analysis with ATPNs. ATPN models for breakdown handling are presented and analyzed for estimating system performance and for designing the optimum number of assembly fixtures. ATPN models can also be used for real-time system control. Finally, possible extensions to this study are discussed.

A Petri net approach to investigating push and pull paradigms in flexible factory automated systems

SUMMARY This paper shows Petri nets (PNs) as a powerful tool to investigate the problem often encountered in manufacturing systems management, namely comparing the performance of a flexible factory automated system (FFAS) operating under push and pull paradigms. The difficulty in solving this type of problems is compounded by many parameters such as processing times at work cells, number of automated guided vehicles and their routeings, lot sizes, and setup times. The PN method to solve such a problem is illustrated by considering an FFAS. Its PN models are formulated and then analysed to compare the performance of FFAS with ‘pushrsquo; and ‘pull’ paradigms. The results show that for the particular FFAS and operational parameters, the ‘push’ paradigm outperforms the ‘pull’ one. It is concluded that PNs are a suitable tool to conduct performance analysis before adopting the ‘push’ or ‘pull’ paradigm.

Real-time Petri nets for modeling, controlling, and simulation of local area networks in flexible manufacturing systems

Abstract Flexible manufacturing systems (FMSs) are advanced manufacturing systems consisting of machining cells, robots, and automated guided vehicles (AGVs) that function under the hierarchical control of computers. The flexibility of an FMS is primarily imparted by integrating the functions of different system elements such as machining cells, robots, and AGVs using computers. Hence, the communication among computers, machining cells, robots, and AGVs in the FMS is important to realize the full benefits of FMSs. Despite the abundant literature on factory communications there is a need for integrated tools to model communication networks in FMS that can be used for both performance evaluation and real-time control. In this paper, the role of local area networks (LANs) in FMSs is explicated. The contribution of this paper is two-fold. With this aim first, the differences between the LANs in conventional computer systems to that of LANs in FMSs are emphasized. Secondly, Real-time PNs are demonstrated as an integrated modeling tool for LAN in FMS. A software package has been developed to simulate the Petri net models (PNMs). The software package developed aids system designers by simulating the PNM and generating certain performance criteria. Furthermore, the same package assists control engineers for the real-time control of the system designed by system designers.

Evaluating the complexity of Petri nets and ladder logic diagrams for sequence controllers design in flexible automation

This paper presents a methodology to evaluate the complexity of Petri nets (PNs) and ladder logic diagrams (LLDs) for sequence controllers design. Complexity of a design is characterized by the number of basic elements used to model the given control logic. More specifically, this paper presents analytical formulas to estimate the number of basic elements to model certain building blocks of logic modeling when PN and LLD are used. Also, by presenting a methodology to use these analytical formulas, this paper precludes the need for physically building the controllers by either PN or LLD for the comparison of their graphical complexity. The results are demonstrated by considering two examples of sequence controllers. Finally, the limitations of the present work are presented along with possible extensions for future research.<<ETX>>

Discrete Event Control Design for Manufacturing Systems Via Ladder Logic Diagrams and Petri Nets: A Comparative Study

Design methods for sequence controllers play a prominent role in advancing industrial automation. The increasing complexity and varying needs of modern discrete manufacturing systems have challenged the traditional design methods such as the use of ladder logic diagrams (LLDs) for programmable logic controllers. The methodologies based on research results in computer science have recently received growing attention by academic researchers and industrial engineers in order to design flexible, reusable, and maintainable control software. Particularly, Petri nets (PNs) are emerging as a very important tool to provide an integrated solution for modeling, analysis, simulation, and control of industrial automated systems. However, in order to establish PNs as alternative to LLDs there is a need for benchmark studies to formally compare them. The contribution of this chapter is two fold. First, certain criteria are identified to compare LLDs and PNs in designing sequence controllers subject to the changing control requirements. The comparison is performed through a practical system after introducing ‘real-time Petri nets’ for discrete-event control purpose. Secondly, some analytical formulas and a methodology are developed to estimate the number of basic elements used in the PN and LLD designs prior to their constructions. The results will be useful for researchers and engineers to design control systems for complex industrial automated systems.

Automatic generation of Petri net models from logic control specifications

There is a growing need to use advanced and formal methodologies for design and synthesis of supervisory controllers for flexible manufacturing systems. This paper presents an algorithm to generate automatically the Petri net (PN) models from logic control specifications. Traditional methods such as ladder logic diagrams and high-level programming languages proved inefficient to develop control software and very difficult to debug and maintain. PNs were demonstrated as an effective tool for logic controller design by several researchers and industrial practitioners. One critical task in this development is to design Petri net models given the sequence control specifications. This paper presents a PN design algorithm which mainly consists in: (1) dividing the sequence of events into groups, (2) building the PN model via top-down refinement, and (3) merging the common places/paths within the model.<<ETX>>

Monitoring of tool wear using artificial neural networks

An online scheme for tool wear monitoring using artificial neural networks (ANNs) has been proposed. Motivated by the fact that the tool wear at a given instance of time depends on the tool wear value at previous instance of time, memory was included in ANN. With this aim, an ANN without memory, an ANN with one phase memory, and an ANN with two are investigated in this study. The advantages and unique characteristics of the proposed tool wear modeling scheme with earlier methods of tool wear estimation are discussed.