Nof S.Y. Springer Handbook of Automation

Подождите немного. Документ загружается.

Material Handling Automation in Production and Warehouse Systems 55.3 Advanced Technologies 975

the general VRP or VSP. However, the problem can be

complicated when the load level of systems has to be

taken into account. A ﬂow cost function determines the

weight value of an edge based on the queue size and

system processing time for one unit in an AMHS,i.e.,

the load level is measured by these factors. It converts

the weight value to a distance value by using the speed

factor of the system. In an actual problem, this task can

be considerably more complicated.

Prioritizing for the unit load is sometimes very use-

ful for vehicle-initiated dispatching rules [55.11]when

the IAMHS becomes a bottleneck in an FMS for a cer-

tain period of time. The priority determined by the

MMS can be used by an AGV controller to determine

a job priority for a vehicle that has just become idle;

for example, the ﬁrst-come ﬁrst-served rule picks up

the job with the longest waiting time for all unit loads

in the queue. If a priority is given to each job from 1

to 5, the priority unit can be treated as a certain time

scale, e.g., 10min, for each unit. Together with the ac-

tual waiting time, the controller can prioritize the unit

loads; for instance, if a unit load waits for a vehicle as-

signment for 5min and its priority given by the MMS

is 3, then its ﬁnal priority can be 10×2+5min,which

is equal to 25min. The prioritizing methods used by the

MMS generally address problems of how to avoid ma-

chine starvation. While various MMS prioritizing rules

can be used based on the constraints of the shop ﬂoor,

the importance of considering bottleneck machines to

determine transfer priorities of unit loads is emphasized

by Koo et al. [55.46]andLi et al. [55.15].

55.3.4 AI and OR Techniques for MHA

It is worthwhile to compare AI search and OR optimiza-

tion techniques with respect to several different criteria

of logical ﬂexibility, computation time, and application

areas in MHA. In general, AI search algorithms deﬁne

problems with four instances: initial state, successor

function, goal test, and path cost function [55.57]. The

initial state is a state in which the given problem starts.

The successor function receives a state as a param-

eter and returns a set of actions and successors. And

the successors are new states reachable from the given

new state. The deﬁnition of the state together with the

successor function is very important to determine the

overall search space of the given problem and informa-

tion necessary for the solution. The goal test determines

whether a given state satisﬁes all the conditions of the

goal state. The path cost function calculates a numerical

cost for each path explored by the successor function.

There are two types of AI search algorithms: unin-

formed and informed search. Uninformed search does

not use prior information to explore a solution. Exam-

ples of uninformed search algorithms are the depth-ﬁrst

search, breadth-ﬁrst search, and bidirectional search.

Informed search utilizes given information for new

states that will be opened during the search. Informed

search is also called heuristic search, which includes

greedy best-ﬁrst search, A* search, memory-bounded,

and local beam search, which again includes simu-

lated annealing, tabu search, and genetic algorithm.

Constraint programming (CP) is one of the AI search

methods that uses a standard structured representation

consisting of the problem domain and constraint.Fig-

ure 55.10 explains the main procedure used by the

ILOG CP solver [55.58]. The domain is a set of pos-

sible values of the variable representing the problem,

and the constraint is a rule that imposes a limitation on

the variable. The most powerful aspect of this method is

that it utilizes the concept of constraint propagation.It

narrowsdown the search space by imposing a constraint

on variables and the constraint imposed further reduces

domains of other variables based on the constraints al-

ready posted on the variables. Among reduced domains

of variables, the method uses a branching process with

a backtracking algorithm to ﬁnd the best solution. Be-

cause of high modeling ﬂexibility, AI techniques have

been popularly used for a wide variety of control appli-

cations such as robotics and automated planning.

The following studies illustrate the use of AI

techniques for MHA. Cho and Egbelu [55.21]use

Decision variables and domains

Initial constraint propagation

Create a search tree

Search strategy

Fail

Backtrack

Constraint propagation

during search

Solution

Search space Constraints

Fig. 55.10 Main solution procedure of CP (courtesy of

ILOG)

Part F 55.3

976 Part F Industrial Automation

Table 55.3 Comparison of AI and OR techniques

Comparison items AI approaches OR approaches Hybrid approaches

Modeling ﬂexibility High Low High

Time horizon Short Long Long

Response time Short Long Medium

Problem size Small Large Large

Illustrations IAMHS design: IAMHS design: AGV vehicle scheduling

AMHS equipment type Performance evaluation [55.25,32], [55.44,60]

selection [55.21] AGV: IAMHS:

AS/RS: Guide path design [55.28], Integrated scheduling

Job sequencing [55.40], AMHS location [55.24,27] of AMHS and FMS [55.52]

Stacker scheduling [55.8],

Rack design [55.26]

AGV:

Deadlock avoidance [55.49]

knowledge-based rules, fuzzy logic, and decision al-

gorithms to address AMHS equipment type selection

problems. Their procedures consist of three phases:

material handling equipment selections for each ma-

terial ﬂow connections, redundancy and excess capac-

ity check, and budget constraint consideration. Chan

et al. [55.59] also solve a similar problem using an ex-

pert system. An order picking sequence problem in an

AS/RS is addressed by Mahajan et al. [55.40]byus-

ing an AI technique. In their procedure, the state is

represented by a sequence of the orders, and a suc-

cess function providing a selection criterion of the

order sequence is developed by a nearest-neighborhood

strategy.

Operations research techniques mainly focus on the

optimization problems based on linear programming

(LP) [55.61]. LP is extended to integer program-

ming (IP) and mixed integer programming (MIP), that

deal with integral variables, quadratic programming

that uses a nonlinear objective function, and nonlin-

ear programming that allows nonlinear functions in

both the constraint and objective function. Stochastic

programming, which incorporates uncertainties in its

modeling, is also a variant of the LP. OR techniques use

well-structured mathematical models of linear, integer,

quadratic or nonlinear models. Simulation and queu-

ing analysis form another important technical area of

stochastic OR, mainly used for performance analysis.

OR techniques are also popular for solving prob-

lems in MHA. Gaskins and Tanchoco [55.28]and

Tanchoco and Sinreich [55.29] formulate AGV guide

path design problems using 0/1 MIP. Nazzal and

McGinnis [55.32] estimate the system capacity require-

ment of the large-scale AMHS in a semiconductor

fab by utilizing a queuing network model. Ting and

Tanchco [55.27]andHuang et al. [55.24] address

the location problems of the AMHSs in facility lay-

outs using MIP formulations. Huang et al. further use

a heuristic approach employing the Lagrangian relax-

ation method.

AI and OR have their backgrounds in computer

science and industrial engineering, respectively. AI

approaches utilize knowledge representation to solve

a problem; however, OR techniques use mathemati-

cal modeling of the problem. Knowledge representation

consists of symbols and mathematical equations with

relationships. OR techniques generally use dedicated

solvers such as CPLEX and LINDO to solve mathemat-

ical models of the problems, whereas AI techniques use

their own languages such as list processing (LISP), and

programming in logics (Prolog). Constraint program-

ming (CP) uses a solver similarly to the OR solvers

but has much greater ﬂexibility in using procedures and

algorithms. A widely known CP solver is the ILOG

solver. One advantage of AI over OR techniques is

their ﬂexibility in expressing problems. Since the AI

techniques listed above do not use strict mathemati-

cal formulations to represent problems, there is a great

deal of ﬂexibility to deal with instances and activities

in the problems. On the other hand, OR techniques

model problems with strict mathematical procedures

that are generally used repeatedly in different problems.

While OR techniques ﬁnd optimal or near-optimal so-

lutions, AI techniques ﬁnd good solutions for the given

problems. OR techniques have focused on large-scale

optimization problems for decision support systems

while AI techniques are rooted in control problems that

have shorter horizons but need reliable solutions. How-

Part F 55.3

Material Handling Automation in Production and Warehouse Systems References 977

ever, it is also true that there has been some overlap

between AI and OR techniques, especially for local

beam search algorithms. Also, a group of researchers

has tried to take advantage of the two techniques by in-

tegratingprocedures inthe techniques [55.44]. For more

complete reviews onthe historyand stateof theart com-

paring AI and OR techniques, refer to Gomes [55.61],

Kobbacy et al. [55.62], and Marcus [55.63].

Table 55.3 compares AI, OR, and their hybrid

approaches with several different characteristics. An

application area that needs great logical ﬂexibility,

such as the selection problem of material handling

systems, tends to use AI techniques more frequently.

Problems with shorter time horizon use AI heuristic

search approaches (second row in the table), and prob-

lems with a longer time horizon tend to use the OR

approaches. Hybrid approaches focus on reducing com-

putation times. OR techniques are frequently used for

AMHS design problems because response time is less

important for them and they consider a large num-

ber of instances in the system. The AGV dispatching

and routing problems tend to use both heuristic and

OR approaches to similar degrees. Approaches inte-

grating AI and OR approaches pursue both ﬂexibility

and optimality and have been applied to very compli-

cated problems for the MHA [55.44], which deal with

AGV scheduling problems and integrated scheduling of

IAMHS with FMS. Examples on application areas and

their approaches used are listed in the last row of Ta-

ble 55.3.

55.4 Conclusions and Emerging Trends

Material handlingautomation (MHA) in productionand

warehouse management systems provides speedy and

reliable infrastructure for information systems in SCM

such as ERP system, FMS, WMS,andMES.Mostof

all, it enhances accurate data tracking during mater-

ial handling in shop ﬂoors and warehouses. Relying

on these accurate data, high-level automation such as

production and procurement planning, scheduling, and

dispatching in the SCM systems can be made much

more reliable; consequently, more intelligent functions

in their decision-making procedures can be added. An-

other trend in MHA, the integrated and automated

material handling system (IAMHS), has been increas-

ingly implemented in various applications to help

ever-complicated material handling operations in large-

scale production and warehouse systems. Theimportant

issues of the IAMHS reviewed in this chapter can be

largely broken down into design and control issues. The

design issues cover material handling equipment selec-

tion, capacity estimation, innovative equipment design,

and system design optimization. The control issues that

have been hard constraints for the higher-level MHA

tend to involve domain-speciﬁc problems for each com-

ponent system in the IAMHS such as the AS/RS,

AGVS, or MMS.

Several potential routes for further increasing the

level of intelligence in the MHA are recognized in

this chapter. While the number of components in the

IAMHS continues to increase, long response time is

regarded as a major limitation during implementa-

tions of new control algorithms. Continuous efforts

to reduce the computation time of algorithms in the

future are desired. It is also pointed out that newly

developed algorithms should take into account system

architecture for practical applications. As another pos-

sibility, a sensor network might be used for diagnosis

of AMHSs since their reliability is becoming critical,

and also its closed feedback mechanism can poten-

tially be used for more precise controls, as seen in each

context.

References

55.1 Material Handling Industry of America

http://www.mhia.org/ir/, (last accessed February

15, 2009)

55.2 J.R. Montoya-Torres: A literature survey on the

design approaches and operational issues of au-

tomated wafer-transport systems for wafer fabs,

Prod. Plan. Control 17(7), 648–663 (2006)

55.3 T. Feare: GM runs in top gear with AS/RS sequenc-

ing, Mod. Mater. Handl. 53(9), 50–52 (1998)

55.4 H.Y.K. Lau, Y. Zhao: Joint scheduling of mater-

ial handling equipment in automated air cargo

terminals, Comput. Ind. 57(5), 398–411 (2006)

55.5 P.J.M. Meersmans, A.P.M. Wagelmans: Dynamic

Scheduling of Handling Equipment at Automated

Part F 55

978 Part F Industrial Automation

Container Terminals, Econometric Institute Report

EI 2001-33 (Erasmus University, Rotterdam 2001)

55.6 Murata Machinery,

http://www.muratec-l-system.com/en/example/

deliver/medical.html (last accessed February 15,

2009)

55.7 L. Kempfer: Produce delivered fresh and fast,

Mater. Handl. Manag. March, 40–42 (2006)

55.8 C.W.R. Lin, Y.Z. Tsao: Dynamic availability-oriented

control of the automated storage/retrieval system.

A computer integrated manufacturing perspective,

Int. J. Adv. Manuf. Technol. 29(9-10), 948–961

(2006)

55.9 T.H. Chang, H.P. Fu, K.Y. Hu: The innovative con-

veying device application for transferring articles

between two-levels of a multi-story building, Int.

J. Adv. Manuf. Technol. 28(1-2), 197–204 (2006)

55.10 B. Rembold, J.M.A. Tanchoco: Modular framework

for the design of material ﬂow systems, Int. J. Prod.

Res. 32(1), 1–21 (1994)

55.11 P.J. Egbelu, J.M.A. Tanchoco: Characterization of

automated guided vehicle dispatching rules, Int.

J. Prod. Res. 22(3), 359–374 (1984)

55.12 L.Hossain,J.D.Patrick,M.A.Rashid:Enterprise

Resource Planning: Global Opportunities and Chal-

lenges (Idea Group, Hershey 2002)

55.13 D.S. Sun, N.S. Park, Y.J. Lee, Y.C. Jang, C.S. Ahn,

T.E. Lee: Integration of lot dispatching and AMHS

control in a 300 mm wafer FAB, IEEE/SEMI Adv.

Semiconduc. Manuf. Conf. Workshop – Adv. Semi-

conduct. Manuf. Excellence (2005) pp. 270–274

55.14 J. Jimenez, B. Kim, J. Fowler, G. Mackulak,

Y.I. Choung, D.J. Kim: Operational modeling and

simulation of an inter-bay AMHS in semiconduc-

tor wafer fabrication, Winter Simul. Conf. Proc. 2,

1377–1382 (2002)

55.15 B. Li, J. Wu, W. Carriker, R. Giddings: Factory

throughput improvements through intelligent in-

tegrated delivery in semiconductor fabrication

facilities, IEEE Trans. Semiconduct. Manuf. 18(1),

222–231 (2005)

55.16 S.S. Garﬁnkel, B. Rosenberg: RFID Applications,

Security, and Privacy (Addison-Wesley, New York

2006)

55.17 K.C. Lee, S. Lee: Integrated network of Proﬁbus-DP

and IEEE 802.11 wireless LAN with hard real-time

requirement, IEEE Int. Symp. Ind. Electron. 3, 1484–

1489 (2001)

55.18 A. Herrera: Wireless I/O devices in process control

systems, Proc. ISA/IEEE Sensors Ind. Conf. (2004)

pp. 146–147

55.19 S. Phoha, T. LaPorta, C. Grifﬁn: Sensor Network

Operations (Wiley, Piscataway 2006)

55.20 I. Walker, A. Hoover, Y. Liu: Handling unpredicted

motion in industrial robot workcells using sensor

networks, Ind. Robot. 33(1), 56–59 (2006)

55.21 C. Cho, P.J. Egbelu: Design of a web-based inte-

grated material handling system for manufacturing

applications, Int. J. Prod. Res. 43

(2), 375–403 (2005)

55.22 G. Nadoli, M. Rangaswami: Integrated modeling

methodology for material handling systems de-

sign, Winter Simul. Conf. Proc. (1993) pp. 785–789

55.23 J.A. Jimenez, G. Mackulak, J. Fowler: Efﬁcient

simulations for capacity analysis and automated

material handling system design in semiconduc-

tor wafer fabs, Winter Simul. Conf. Proc. (2005)

pp. 2157–2161

55.24 S. Huang, R. Batta, R. Nagi: Variable capacity sizing

and selection of connections in a facility layout, IIE

Trans. 35(1), 49–59 (2003)

55.25 Y.J. Jang, G.H. Choi, S.I. Kim: Modeling and analysis

of stocker system in semiconductor and LCD fab,

IEEEInt.Symp.Semiconduct.Manuf.Conf.Proc.

ISSM 2005 (2005) pp. 273–276

55.26 Y.H. Lee, M.H. Lee, S. Hur: Optimal design of rack

structure with modular cell in AS/RS, Int. J. Prod.

Econ. 98(2), 172–178 (2005)

55.27 J.-H. Ting, J.M.A. Tanchoco: Optimal bidirectional

spine layout for overhead material handling sys-

tems, IEEE Trans. Semiconduct. Manuf. 14(1), 57–64

(2001)

55.28 R.J. Gaskins, J.M.A. Tanchoco: Flow path design for

automated guided vehicle systems, Int. J. Prod.

Res. 25(5), 667–676 (1987)

55.29 J.M.A. Tanchoco, D. Sinriech: OSL – optimal single

loop guide paths for AGVS, Int. J. Prod. Res. 30(3),

665–681 (1992)

55.30 Y.A. Bozer, M.M. Srinivasan: Tandem AGV system:

a partitioning algorithm and performance compar-

ison with conventional AGV systems, Eur. J. Oper.

Res. 63, 173–191 (1992)

55.31 P. Caricato, A. Grieco: Using simulated annealing

to design a material-handling system, IEEE Intell.

Syst. 20(4), 26–30 (2005)

55.32 D. Nazzal, L.F. McGinnis: Analytical approach to es-

timating AMHS performance in 300 mm fabs, Int. J.

Prod. Res. 45(3), 571–590 (2007)

55.33 I.F.A. Vis, R. de Koster, K.J. Roodbergen,

L.W.P. Peeters: Determination of the number of

automatedguidedvehiclesrequiredatasemi-

automated container terminal, J. Oper. Res. Soc.

52(4), 409–417 (2001)

55.34 B.A. Peters, T. Yang: Integrated facility layout and

material handling system design in semiconduc-

tor fabrication facilities, IEEE Trans. Semiconduct.

Manuf. 10(3), 360–369 (1997)

55.35 J. Chung, J. Jang: The integrated room layout for

semiconductor facility plan, IEEE Trans. Semicon-

duct. Manuf. 20(4), 517–527 (2007)

55.36 B. Rembold, J.M.A. Tanchoco: Material ﬂow system

model evaluation and improvement, Int. J. Prod.

Res. 32(11), 2585–2602 (1994)

Part F 55

Material Handling Automation in Production and Warehouse Systems References 979

55.37 I.F.A. Vis: Survey of research in the design and con-

trol of automated guided vehicle systems, Eur. J.

Oper. Res. 170(3), 677–709 (2006)

55.38 T. Le-Anh, M.B.M. De Koster: A review of design

and control of automated guided vehicle systems,

Eur. J. Oper. Res. 171(1), 1–23 (2006)

55.39 M. Dotoli, M.P. Fanti: A coloured Petri net model for

automated storage and retrieval systems serviced

by rail-guided vehicles: a control perspective, Int.

J. Comput. Int. Manuf. 18(2-3), 122–136 (2005)

55.40 S. Mahajan, B.V. Rao, B.A. Peters: A retrieval

sequencing heuristics for miniload end-of-aisle

automated storage/retrieval system, Int. J. Prod.

Res. 36(6), 1715–1731 (1998)

55.41 C. Lee, B. Liu, H.C. Huang, Z. Xu, P. Golds-

man: Reservation storage policy for AS/RS at air

cargo terminals, Winter Simul. Conf. Proc. (2005)

pp. 1627–1632

55.42 O.V.K. Chetty, M.S. Reddy: Genetic algorithms for

studies on AS/RS integrated with machines, Int. J.

Adv. Manuf. Technol. 22(11-12), 932–940 (2003)

55.43 D. Sinriech, L. Palni: Scheduling pickup and

deliveries in a multiple-load discrete carrier envi-

ronment, IIE Trans. Inst. Ind. Eng. 30(11), 1035–1047

(1998)

55.44 A.I. Corréa, A. Langevin, L.M. Rousseau: Scheduling

and routing of automated guided vehicles: a hy-

brid approach, Comput. Oper. Res. 34(6), 1688–1707

(2007)

55.45 J. Jang, J. Suh, P.M. Ferreira: An AGV routing policy

reﬂecting the current and future state of semicon-

ductor and LCD production lines, Int. J. Prod. Res.

39(17), 3901–3921 (2001)

55.46 P.H. Koo, J. Jang, J. Suh: Vehicle dispatching for

highly loaded semiconductor production consider-

ing bottleneck machines ﬁrst, Int. J. Flex. Manuf.

Syst. 17(1), 23–38 (2005)

55.47 C.W. Kim, J.M.A. Tanchoco, P.-H. Koo: AGV dis-

patching based on workload balancing, Int. J.

Prod. Res. 37(17), 4053–4066 (1999)

55.48 B.H. Jeong, S.U. Randhawa: A multi-attribute

dispatching rule for automated guided vehicle sys-

tems,Int.J.Prod.Res.39(13), 2817–2832 (2001)

55.49 R.L. Moorthy, W. Hock–Guan, W.-C. Ng, T. Chung–

Piaw: Cycle deadlock prediction and avoidance for

zone controlled AGV system, Int. J. Prod. Econ. 83,

309–324 (2003)

55.50 G. Bruno, G. Ghiani, G. Improta: Dynamic position-

ing of idle automated guided vehicles, J. Intell.

Manuf. 11(2), 209–215 (2000)

55.51 R. Cheung, A. Lee, D. Mo: Flow diversion ap-

proaches for shipment routing in automatic

shipment handling systems, Proc. – IEEE Int. Conf.

Robot. Autom. (2006) pp. 695–700

55.52 S. Sujono, R.S. Lashkari: A multi-objective model of

operation allocation and material handling system

selection in FMS design, Int. J. Prod. Econ. 105(1),

116–133 (2007)

55.53 J. Paulo, R.S. Lashkari, S.P. Dutta: Operation al-

location and materials-handling system selection

in a ﬂexible manufacturing system: a sequential

modeling approach, Int. J. Prod. Res. 40,7–35

(2002)

55.54 R.S. Lashkari, R. Boparai, J. Paulo: Towards an

integrated model of operation allocation and ma-

terials handling selection in cellular manufacturing

system, Int. J. Prod. Econ. 87(2), 115–139 (2004)

55.55 R.U. Ayres: Complexity, reliability and design:

manufacturing implications, Manuf. Rev. 1(1), 26–

35 (1988)

55.56 R.K. Ahuja, T.L. Magnanti, J.B. Orlin: Network

ﬂows: theory, algorithms, and applications (Pren-

tice Hall, Upper Saddle River 1993)

55.57 S. Russell, P. Norvig: Artiﬁcial Intelligence: a Mod-

ern Approach (Prentice Hall, New York 2003)

55.58 ILOG Solver 5.3 user manual

55.59 F.T.S. Chan, R.W.L. Ip, H. Lau: Integration of ex-

pert system with analytic hierarchy process for the

design of material handling equipment selection

system, J. Mater. Process. Technol. 116(2-3), 137–145

(2001)

55.60 D. Naso, B. Turchiano: Multicriteria meta-heuristics

for AGV dispatching control based on computa-

tional intelligence, IEEE Trans. Syst. Man Cybern.

B 35(2), 208–226 (2005)

55.61 C.P. Gomes: Artiﬁcial intelligence and operations

research: challenges and opportunities in planning

and scheduling, Knowl. Eng. Rev. 15(1), 1–10 (2000)

55.62 K.A.H. Kobbacy, S. Vadera, M.H. Rasmy: AI and or

in management of operations: history and trends,

J. Oper. Res. Soc. 58(1), 10–28 (2007)

55.63 R. Marcus: Application of artiﬁcial intelligence to

operations research, Commun. ACM 27(10), 1044–

1047 (1984)

Part F 55

“This page left intentionally blank.”

981

Industrial Co

56. Industrial Communication Protocols

Carlos E. Pereira, Peter Neumann

This chapter discusses a very relevant aspect in

modern automation systems: the presence of

industrial communication networks and their pro-

tocols. The introduction of Fieldbus systems has

been associated with a change of paradigm to de-

ploy distributed industrial automation systems,

emphasizing device autonomy and decentralized

decision making and control loops. The chapter

presents the main wired and wireless industrial

protocols used in industrial automation, man-

ufacturing, and process control applications. In

order to help readers to better understand the

differences between industrial communication

protocols and protocols used in general computer

networking, the chapter also discusses the spe-

ciﬁc requirements of industrial applications. As the

trend of future automation systems is to incorpo-

rate complex heterogeneous networks, consisting

of (partially homogeneous) local and wide area as

well as wired and wireless communication sys-

tems, the concept of virtual automation networks

is presented

56.1 Basic Information ................................. 981

56.1.1 History...................................... 981

56.1.2 Classiﬁcation ............................. 982

56.1.3 Requirements in Industrial

Automation Networks ................. 982

56.1.4 Chapter Overview ....................... 982

56.2 Virtual Automation Networks................. 983

56.2.1 Deﬁnition, Characterization,

Architectures ............................. 983

56.2.2 Domains ................................... 983

56.2.3 Interfaces, Network Transitions,

Transmission Technologies .......... 984

56.3 Wired Industrial Communications .......... 984

56.3.1 Introduction .............................. 984

56.3.2 Sensor/Actuator Networks............ 985

56.3.3 Fieldbus Systems........................ 986

56.3.4 Controller Networks .................... 988

56.4 Wireless Industrial Communications ....... 991

56.4.1 Basic Standards ......................... 991

56.4.2 Wireless Local Area Networks

(WLAN) ...................................... 992

56.4.3 Wireless Sensor/Actuator

Networks .................................. 992

56.5 Wide Area Communications ................... 993

56.6 Conclusions.......................................... 995

56.7 Emerging Trends .................................. 995

56.8 Further Reading ................................... 997

56.8.1 Books ....................................... 997

56.8.2 Various Communication

Standards ................................. 997

56.8.3 Various web Sites

of Fieldbus Organizations

and Wireless Alliances ................ 997

References .................................................. 998

56.1 Basic Information

56.1.1 History

Digital communication is now well established in dis-

tributed computer control systems both in discrete

manufacturing as well as in the process control in-

dustries. Proprietary communication systems within

SCADA (supervisory control and data acquisition) sys-

tems have been supplemented and partially displaced

by Fieldbus and sensor bus systems. The introduction

of Fieldbus systems has been associated with a change

of paradigm to deploy distributed industrial automation

systems, emphasizing device autonomy and decentral-

Part F 56

982 Part F Industrial Automation

ized decision making and control loops. Nowadays,

(wired) Fieldbus systems are standardized and are the

most important communication systems used in com-

mercial control installations. At the same time, Ethernet

won the battle as the most commonly used communi-

cation technology within the ofﬁce domain, resulting in

low component prices caused by mass production. This

has led to an increasing interest in adapting Ethernet

for industrial applications and several approaches have

been proposed (Sect. 56.1.4). Ethernet-based solutions

are dominating as a merging technology.

In parallel to advances on Ethernet-based industrial

protocols, the use of wireless technologies in the in-

dustrial domain has also been increasingly researched.

Following the trend to merge automation and ofﬁce

networks, heterogeneous networks (virtual automation

networks (VA N)), consisting of local and wide area net-

works, as well as wired and wireless communication

systems, are becoming important [56.1].

56.1.2 Classiﬁcation

Industrial communication systems can be classiﬁed as

follows regarding different capabilities:

•

Real-time behavior: Within the automation domain,

real-time requirements are of uttermost importance

and are focused on the response time behavior of

data packets. Three real-time classes can be identi-

ﬁed based on the required temporal behavior:

– Class 1: soft real-time. Scalable cycle time, used

in factory ﬂoor and process automation in cases

where no severe problems occur when deadlines

are not met.

– Class 2: hard real-time. Typical cycletimes from

1 to 10 ms, used for time-critical closed loop

control.

– Class 3: isochronous real-time, cycle times from

250μs to 1 ms, with tight restrictions on jitter

(usually less than 1μs), used for motion control

applications.

Additionally, there is a class non real-time,which

means systems without real-time requirements;

these are not considered here. It means (regarding

industrial automation) exchange of engineering data

maintenance, etc.

•

Distribution: The most important achievement

of industrial communication systems are local

area communication systems, consisting of sen-

sor/actuator networks (Chap.20 and Sect.56.3.2),

Fieldbussystems, and Ethernet-based local area net-

works (LAN). Of increasing importance is the use

of wide area networks (WA N) (telecommunication

networks, Internet, etc.). Thus, it should be advan-

tageous to consider WA N s as part of an industrial

communication system (Sect.56.2), mostly within

the upper layers of an enterprise hierarchy.

•

Homogeneity: There are homogeneous parts (e.g.

standardized Fieldbus systems) within an industrial

communication system. But in real applications the

use of heterogeneous networks is more common,

especially when using WA Ns and when connected

with services of network providers.

•

Installations types: While most of the installed en-

terprise networks are currently wired, the number

of wireless installations is increasing and this trend

will continue.

56.1.3 Requirements in Industrial

Automation Networks

The main requirements are:

•

Real-time behavior: Diagnosis, maintenance, com-

missioning, and slow mobile applications are

examples of non real-time applications. Process

automation and data acquisition usually present

soft real-time requirements. Examples of hard

real-time applications are closed-loop control ap-

plications, such as in fast mobile applications and

machine tools. Motion control is an example of an

isochronous hard real-time application.

•

Functionalsafety: Protection against hazards caused

by incorrect functioning including communica-

tion via heterogeneous networks. There are several

safety integrity levels (SIL) [56.2]. It includes the

inﬂuence of noisy environments and the degree of

reliability.

•

Security: This means a common security concept

for distributed automation using a heterogeneous

network with different security integrity levels (not

existent yet).

•

Location awareness: The desired context awareness

leads to the usage of location-based communication

services and context-sensitive applications.

56.1.4 Chapter Overview

The remainder of the chapter is structured as follows.

Section 56.2discusses the concept of virtual automation

networks (VANs), a key concept in future distributed

automation systems, which will be composed of (par-

tially homogeneous) local and wide area as well as

Part F 56.1

Industrial Communication Protocols 56.2 Virtual Automation Networks 983

wired and wireless communication systems leading to

complex heterogenous communication networks.

In Sect.56.3 the main wired industrial communi-

cation protocols are presented and compared, while

Sect.56.4 discusses wireless industrial communication

systems. Section 56.5 deals with the use of wide area

communications to execute remote automation opera-

tions.

56.2 Virtual Automation Networks

56.2.1 Deﬁnition, Characterization,

Architectures

Future scenarios of distributed automation lead to

desired mechanisms for geographically distributed au-

tomation functions for various reasons:

•

Centralized supervision and control of (many) de-

centralized (small) technological plants

•

Remote control, commissioning, parameterization,

and maintenance of distributed automation systems

•

Inclusion of remote experts or external machine-

readable knowledge for plant operation and main-

tenance (for example, asset management, condition

monitoring, etc.).

This means that heterogeneous networks, consist-

ing of (partially homogeneous) local and wide ar-

eas, as well as wired and wireless communication

systems, will play an increasing role. Figure 56.1

Industrial domain

Industrial segment

Domain connected

via radio link

Public and private

telecommunication

networks/internet

ain

Indust

VAN domain A

onnected

dio link

VAN domain B

Public and private

telecommunicati

networks/inte

VAN domain C

Intrinsic safety

domain

Office sub domains

Office domain

Individual industrial sub domains

Remark:

All systems are shown generally as a bus. Depending on the real system

it may be any type of topology including a ring.

Real-time domain

Single device integration

(e.g. tele control)

Industrial WLAN domain

Industrial backbone

Mobile devices

Industrial

WLAN domain

Real-time domain

Mobile devices

Remote industrial domains / subsidiary / customer sites

Fig. 56.1 Different VA N domains related to different automation applications [56.3]

depicts the communication environment of a com-

plex automation scenario. Following a unique design

concept, regarding the objects to be transmitted be-

tween geographically distributed communication end

points, the heterogeneous network becomes a virtual

automation network (VAN) [56.3, 4]. VAN charac-

teristics are deﬁned for domains, where the expres-

sion domain is widely used to address areas and

devices with common properties/behavior, common

network technology, or common application pur-

poses.

56.2.2 Domains

Within the overall automation and communication en-

vironment, a VAN domain covers all devices that are

grouped together on a logical or virtual basis to rep-

resent a complex application such as an industrial

application. Therefore, the encompassed networks may

be heterogeneous and devices can be geographically

Part F 56.2

984 Part F Industrial Automation

W/WL

link

W/WL

link

device

MAN

station

WAN,

Internet

Device

Con-

troller

Super-

visor

Internet

server

Office

device

Manager

station

Telecommunication

LAN

RTE

Fieldbus

Wireless

fieldbus

Proxy

gateway

Fig. 56.2 Network transitions (local area networks (LAN), wire-

less LAN (WL), wired/wireless (W/WL), real-time Ethernet (RTE),

metropolitan area network (MAN), wide area network (WAN ))

distributed over a physical environment, which shall

be covered by the overall application. But all de-

vices that have to exchange information within the

scope of the application (equal to a VA N domain)

must be VAN aware or VAN enabled devices. Other-

wise, they are VAN independent and are not a member

of a VAN domain. Figure 56.1 depicts VAN domain

examples representing three different distributed appli-

cations.

Devices related to a VAN domain may reside in

a homogeneous network domain (e.g. the industrial

domain showninFig.56.1). But, depending on the ap-

plication, additional VA N relevant devices may only be

reached by crossing other network types (e.g.,wide area

network type communication) or they need to use proxy

technology to be represented in the VAN domain view

of a complex application.

56.2.3 Interfaces, Network Transitions,

Transmission Technologies

A VAN network consists of several different com-

munication paths and network transitions. Figure 56.2

depicts the required transitions in heterogeneous net-

works.

Depending on the network and communication

technology of the single path there will be differences

in the addressing concept of the connected network seg-

ments. Also the communication paths have different

communication line properties and capabilities. There-

fore, for the path of two connected devices within

a VA N domain the following views are possible:

•

The logical view: describing the properties/capabili-

ties of the whole communication path

•

The physical view: describing the detailed proper-

ties/capabilities of the passed technology-dependent

communication paths

•

The behavioral view: describing the different

cyclic/acyclic temporal behavior of the passed seg-

ments.

There are different opportunities to achieve a com-

munication path between network segments/devices

(or their combinations). These are: Ethernet line

(with/without transparent communication devices),

wireless path, telecommunication network (1:1), public

networks (n:m, provider-oriented), VPN tunnel, gate-

way (without application data mapping), proxy (with

application data mapping), VAN access point, and IP

mapping. All networks, which can not be connected via

an IP-based communication stack, must be connected

using a proxy. For connecting nonnested/cascaded VA N

subdomains via public networks the last solution (VAN

access point) should be preferred.

56.3 Wired Industrial Communications

56.3.1 Introduction

Wired digital communication has been an important

driving force of computer control systems for the last

30years. To allow the access to data in various layers

of an enterprise information system by different users,

there is a need to merge different digital communica-

tion systems within the plant, control, and device levels

of an enterprisenetwork. On these different levels, there

are distinct requirements dictated by the nature and type

of information being exchanged. Network physical size,

number of supported devices, network bandwidth, re-

sponse time, sampling frequency, and payload size are

some of the performance characteristics used to clas-

sify and group speciﬁc network technologies. Real-time

requirements depend on the type of messages to be ex-

changed: deadlines for end-to-end data transmission,

maximum allowed jitter for audio and video stream

transmission, etc. Additionally, available resources at

the various network levels may vary signiﬁcantly. At

Part F 56.3