Nof S.Y. Springer Handbook of Automation

Подождите немного. Документ загружается.

Welding Automation 59.5 Robotic Welding 1035

1. Approaching the workpiece without weaving, de-

tecting electrical contact between the electrode and

the weld plates, and calculating the starting position

from this information

2. Approaching the workpiece with weaving, detecting

the arc current, and tracking the seam in a normal

way

For welding automation, other sensors can also be

used, for example: inductive proximity sensors or eddy-

current sensors,infrared sensors,ultrasonic sensors, and

also sophisticated computer vision. However, they are

not so common in this application, which is charac-

terized by harsh environment with high temperatures,

intense light, and high currents.

59.4.2 Monitoring and Control of Welding

Using data obtained from the sensors, it is possible to

evaluate the weld quality and detect (or even classify)

different weld defects, such as porosity, metal spatter,

irregular bead shape, incomplete penetration, etc. These

capabilities are implemented in monitoring systems,

which usually use high-speed online analysis of weld-

ing voltage and/or current that are compared with preset

nominal values or time patterns. Based on this analysis,

an alarm is triggered if any difference from the preset

values exceeds the given threshold. More sophisticated

installations use computer-based image processing to

evaluate the welding pool geometry and penetration

depth [59.11].

To judge weld quality, the monitoring system relies

on physical or statistical models, allowing the deﬁnition

of alarm thresholds correlated with real weld defects

or welding process speciﬁcations; for instance, for all

GMAW processes, the voltage and current shape and

mean values allows the detection of the metal trans-

fer mode (short circuit, globular or spray transfer). In

pulsed GMAW, the peak current is monitored and com-

pared with preset values. For short-circuit GMAW,the

monitoring features includes short-circuit time or fre-

quency, as well as the average short-circuit current and

the average arc current. In the general case, the features

used for monitoring may be dependent on the speciﬁc

algorithm and the welding condition.

For process feature analysis, various strategies are

applied. The simplest ones employ deterministic deci-

sion making based on nominal values and tolerances,

where any deviation from these is considered a po-

tential cause of quality decrease. More sophisticated

techniques employ template matching or treat the meas-

ured features as random variables and apply statistical

methods such as control charts or spectrum analy-

sis. However, the user must realize that increasing

the detection probability often leads to false alarms

that regularity interrupt the process. So, most current

commercial monitoring systems utilize simple and ro-

bust algorithms, in which process features are averaged

within user-deﬁned time segments, ﬁltered, and com-

pared with a predeﬁned threshold corresponding to

normal welding conditions.

Welding process deviations detected via monitor-

ing are to be compensated by control actions [59.12].

However, because of process complexity and indirect

relevance of the observable data, simple feedback loops

cannot be implemented. So, in addition to seam track-

ing, model-based strategies must be applied to enable

adjustment of thewelding equipmentsettings. However,

in spite of its tremendous practical signiﬁcance, this is

still an active research area that employs various sophis-

ticated decision-making techniques based on artiﬁcial

intelligence and knowledge-based modeling.

59.5 Robotic Welding

Most industrial automated welding systems employ

robotic manipulators, which are integrated with stan-

dard welding equipment that provides energy supply

and basic control of welding parameters. The manip-

ulators replace the human operator by handling the

welding tool and positioning the workpiece. Usually

this leads to an increase in quality and productivity,

but poses a number of additional problems related to

robot control, programming, calibration, and mainte-

nance [59.13,14].

59.5.1 Composition

of Welding Robotic System

Currently, robots are mainly used for arc and spot weld-

ing processes. However, some recent applications deal

with laser and plasma welding and also with friction stir

welding. Typically, a robotic welding station includes

a robot, a robot controller, welding equipment with rel-

evant sensors, and clamping devices (ﬁxture), allowing

the workpiece to be held in the desired position in spite

Part F 59.5

1036 Part F Industrial Automation

a) b)

Fig. 59.8a,b Composition of arc welding robotic stations with ﬂoor-mounted (a) and column-mounted (b) robots (af-

ter [59.15, pp. 595, 602])

of thermal deformations. In addition, there are a vari-

ety of auxiliary mechanisms that provide an increasing

in the robot workspace, better weld positioning, safety

protection, and workpiece transportation betweenwork-

stations.

The most common design of an arc welding robotic

station is shown in Fig.59.8. Usually, the robot has

six actuated axes, so it can access any point within the

working range at any orientation of the welding torch.

In most cases robots are implemented with a serial

architecture with revolute joints, which ensures larger

workspaces (Fig.59.9). Typical arc welding robots have

a working envelope of 2000mm and payload capacity

of about 5 kg, which is sufﬁcient for handling weld-

ing tools. To extend the working range, robots may

be installed in an overhead position. A further exten-

sion of the working range can be achieved by installing

the robot onto a linear carriage with auxiliary actu-

ated axes (track, gantry or column). The wire feed unit

and the spool carriers for the wire electrodes are often

ﬁxed to the robot, but can also be placed separately.

In many cases the torch is equipped with shock ab-

a) b) c) d)

Fig. 59.9a–d Mechanical components of an arc welding robotic station: (a) robotic manipulator, (b) positioner;

(c),(d) robots with translational motion units (http://www.kuka.com)

sorption devices (such as springs) to protect it against

collisions.

The workpiece positioners allow the location of

seams in the bestposition relative to gravity (i. e., down-

hand) and to provide better weld accessibility. They

usually have one or two actuated axes and may han-

dle payload from a few kilograms to several hundred

tons. The most common positioners are turnover, turn–

tilt,andorbital tables,butturning rolls are also used

to rotate the workpiece while making circular seams (in

tank manufacturing, for instance). Positioners with or-

bital design have an advantage for heavy parts, allowing

rotation of the workpiece around its center of gravity. In

some cases, positioners are implemented with a mul-

titable architecture, in which the operator feeds and

removes the welded workpiece on one side, while the

robot is welding on the other side. The positioner axes

may either turn to certaindeﬁned positions(index-based

control) or be guided by the robot controller and moved

synchronically with the internal axes.

For spot welding, the robot payload capacity is es-

sentially higher (about 150kg), being deﬁned by rather

Part F 59.5

Welding Automation 59.5 Robotic Welding 1037

Fig. 59.10 Robotic spotwelding linefor the automotivein-

dustry (http://tal.co.in/solutions/equipments/robotics-

automation.html)

heavy equipment mounted on the robotic manipulator

arm. Usually, each spot welding station includes several

robots working simultaneously to provide the same cy-

cle time along the manufacturing line. An example of

a spot welding line for car manufacturing is presented

in Fig.59.10. For such applications, robots usually per-

form several thousands welds on over a hundred parts

with a cycle timeof about 1min. Besidesthe joiningand

handling operations, the robots also ensure online mea-

surement and inspection by means of dedicated laser

sensors.

To ensure coordination of all components of the

automated welding system, a relevant multimicropro-

cessor control architecture usually comprises two hier-

archical levels. At the lower level, the local controllers

implement mainly position-based algorithms that can

receive a desired trajectory and run it continuously

(for each actuated axis separately, but simultaneously

and coordinated). High-level controllers ensure tra-

jectory correction in real time, as a function of the

observed results of the welding process. Some robot

controllers can be connected via the Internet with tele-

diagnostic systems to support service personnel during

troubleshooting.

59.5.2 Programming of Welding Robots

To take advantage of robotic welding, especially in

small-batch manufacturing, it is necessary to reduce the

prewelding phase (or setup time), which includes se-

lection of welding parameters and generation of control

program deﬁning motions of the robot, positioner, and

other related mechanisms. This process is time con-

suming and may be longer than the actual welding

phase [59.16].

For selection of welding parameters, there are at

present a number of generally accepted databases. They

allow the deﬁnition of optimal values of the weld-

ing current, voltage, welding speed, wire diameter, and

number of weld beads/layers depending on type of

weld, welding position, properties of materials, plate

thickness, etc. Also, these databases usually provide in-

terface to computer-aided design (CAD) models of the

joining components to simplify extraction of geometri-

cal information.

For robot programming, two basic methods exist:

online (programming at the robot) and ofﬂine (program-

ming out of the robot cell). The former method, which

is also referred to as manual teaching, requires extrac-

tion of the robot from the manufacturing process and

involves operator-guided implementation of all required

motions. The operator uses a dedicated teach-pendant

to move the welding torch to notable points of the

weld, to store the torch position and orientation, and

to create corresponding motion commands with neces-

sary attributes (deﬁning velocity, type of interpolation

along the path,weave pattern,welding parameters,etc.).

The simplest implementation of the ofﬂine program-

ming uses an external computer to create a text ﬁle

describing the sequence of motions, but the command

arguments (i. e., torch position and orientation) are ob-

tained via manual teaching. Nevertheless, this offer

essential shortening of programming time because of

extensive use of standard macros.

Advanced ofﬂine programming systems provide

fully autonomous program generation, completely out-

side of the manufacturing cell. They rely on so-

Fig. 59.11 Simulation and programming environment of

eM-Workplace (Robcad) (http://www.ugs.com)

Part F 59.5

1038 Part F Industrial Automation

phisticated photorealistic 3-D graphical simulation of

the robotic system and parts to be welded, allowing

the required torch coordinates/orientations to be ob-

tained directly from the models. Moreover, modern

CAD-based robotic programming systems [such as em-

Workplace (Robcad), IGRIP, CimStation, etc.] provide

an interface to all standard 3-D modeling systems and

incorporate a number of additional tools for robotic

cell design, layout optimization, graphical simulation

of the movements, program debugging and veriﬁca-

tion with respect to collisions and cycle time, program

downloading to the robot controller, unloading of ex-

isting programs for optimization, etc. An example view

of a CAD-based simulation and programming environ-

ment is presented in Fig.59.11. However, while using

the ofﬂine programming, it is necessary to ensure good

correspondence between the nominal CAD model of the

robot and its actual geometrical parameters. In prac-

tice, this is not a trivial problem, which is solved via

calibration of all geometrical parameters describing the

workcell components and their spatial location.

Automation of robot programming is still an active

research area, which is targeted to replace movement-

oriented program development to task-oriented pro-

gramming. The ﬁnal goal is automatic generation of

robot programs from CAD drawings and welding

databases, similar to programming methods for com-

puter numerical control (CNC) machines.

59.6 Future Trends in Automated Welding

At present, welding automation is on the rise because

of stricter customer demands and a shortage of skilled

welders. So, equipment manufactures and system inte-

grators are enhancing their production and implement-

ing more advanced technologies. The most important

technology-oriented directions deﬁning future trends in

welding automation are the following [59.17–20]:

•

Improvement of traditional welding processes with

respect to productivity andenvironmental issues (in-

cluding development of better controllable power

sources, new electrodes, shielding gases and ﬂaxes;

using twin-arc and tandem-arc torches)

•

Industrial implementation of new efﬁcient and

environment-friendly processes, such as laser beam

and electron beam welding, friction stir welding,

and magnetic pulse welding (includingdevelopment

of new energy sources, relevant control algorithms,

manipulating equipment)

•

Creating new knowledge-based welding process

models with ability for online learning and ca-

pability for online feedback control of essential

process features (such as molten pool geometry,

heat distribution, surface temperature proﬁle, ther-

mal deformations)

•

Development of advanced sensors and intelligent

seam-tracking control algorithms (to compensate

for parts’ mechanical tolerances in 3-D or 6-D

space, and making welds in nonﬂat positions)

•

Development of new process monitoring and non-

destructive evaluation methods (using model-based

condition monitoring and failure analysis tech-

niques, and online ultrasonic and laser-based test-

ing)

Concerning mechanical components (robotic ma-

nipulators, positioners, etc.), it is recognized that their

current performance satisﬁes the requirements of most

welding processes with respect to ability to reproduce

the desired trajectory with given speed and accuracy.

Future developments will focus on automation of robot

programming and integration with other equipment:

•

Task-oriented ofﬂine programming and integra-

tion with product design (using simulation-based

methods; simultaneous product and ﬁxture de-

sign; implementing 3-D virtual-reality tools and

concurrent engineering concepts; developments of

human–machine interfaces)

•

Standardization of mechanical components, control

platforms, sensing devices and control architectures

(to reduce the system development time/cost and

simplify its modiﬁcation)

•

Monitoring of the welding equipment and robotic

manipulators (to predict or detect machine failures

and reduce downtime using predeﬁned exception-

handling strategies)

In addition, there are a number of essential issues

that are not directly linked with welding technol-

ogy and equipment. These include marketing aspects,

and also networking and collaboration using modern

e-Manufacturing concepts.

Part F 59.6

Welding Automation References 1039

59.7 Further Reading

•

AMS: ASMHandbook. Vol. 6: Welding, Brazing and

Soldering, 10th edn. (ASM International, Metals

Park 1993)

•

K. Weman: Welding Processes Handbook (Wood-

head, Boca Raton 2003)

•

J. Pan: Arc Welding Control (CRC, Boca Raton

2003)

•

R. Hancock, M. Johnsen:Developments in guns and

torches, Weld. J. 83(5), 29–32 (2004)

•

F.M. Hassenkhoshnaw, I.A. Hamakhan: Automa-

tion capabilities for TIG and MIG welding pro-

cesses, Weld. Cutt. 5(3), 154–156 (2006)

•

M. Fridenfalk, G. Bolmsjö: Design and validation

of a universal 6-D seam tracking system in robotic

welding based on laser scanning, Ind. Robot 30(5),

437–448 (2003)

•

A.P. Pashkevich, A. Dolgui: Kinematic control of a

robot-positioner system for arc welding application.

In: Industrial Robotics: Programming, Simulation

and Applications, ed. by K.-H. Low (Pro Literatur,

Mammendorf 2007) pp.293–314

•

G.E. Cook, R. Crawford, D.E. Clark, A.M. Strauss:

Robotic friction stir welding, Ind. Robot 31(1), 55–

63 (2004)

•

X. Chen, R. Devanathan, A.M. Fong (eds.): Ad-

vanced Automation Techniques in Adaptive Mater-

ial Processing (World Scientiﬁc, River Edge 2002)

•

U. Dilthey, L. Stein, K. Wöste, F. Reich: De-

velopments in the ﬁeld of arc and beam weld-

ing processes, Weld. Res. Abroad 49(10), 21–31

(2003)

•

G. Bolmsjö, M. Olsson, P. Cederberg: Robotic

arc welding – trends and developments for

higher autonomy, Ind. Robot 29(2), 98–104

(2002)

•

J. Villafuerte: Advances in robotic welding technol-

ogy, Weld. J. 84(1), 28–33 (2005)

•

T. Yagi: State-of-the-art welding and de-burring

robots, Ind. Robot 31(1), 48–54 (2004)

•

A. Benatar, D.A. Grewell, J.B. Park (eds.): Plastics

and Composites Welding Handbook (Hanser Gard-

ner, Cincinnati 2003)

•

W. Zhang: Recent advances and improvements in

the simulation of resistance welding processes,

Weld. World 50(3–4), 29–37 (2006)

•

P.G. Ranky: Reconﬁgurable robot tool designs and

integration applications, Ind. Robot 30(4), 38–344

(2003)

•

W.G. Rippey: Network communications for weld

cell integration – status of standards development,

Ind. Robot 31(1), 64–70 (2004)

•

The Welding Institute, UK. http://www.twi.co.uk

•

American Welding Society, USA.

http://www.aws.org

•

ASM International: Materials Information Society.

http://www.asminternational.org

•

M. Sciaky: Spot welding and laser welding. In:

Handbook of Industrial Robotics, ed. by S.Y. Nof

(Wiley, New York 1999) pp.867–886

•

J.A. Ceroni: Arc welding. In: Handbook of Indus-

trial Robotics, ed. by S.Y. Nof (Wiley, New York

1999) pp.887–905

References

59.1 L.F. Jeffus: Welding: Principles and Applications,

6th edn. (Delmar, New York 2007)

59.2 H.B. Cary, S.C. Helzer: Modern Welding Technology,

6th edn. (Prentice Hall, New Jersey 2004)

59.3 W.A. Bowditch, K.E. Bowditch, M.A. Bowditch:

Welding Technology Fundamentals (Goodheart-

Willcox, South Holland 2005)

59.4 AWS: AWS Welding Handbook. Vol. 1: Welding Sci-

ence and Technology. Vol. 2: Welding Processes,

9th edn. (American Welding Society, Miami 2001)

59.5 H.B. Cary: Arc Welding Automation (Marcel Dekker,

New York 1995)

59.6 J. Norrish: Advanced Welding Processes (IOP, Lon-

don 2006)

59.7 G. Bolmsjö, M. Olsson: Sensors in robotic arc weld-

ing to support small series production, Ind. Robot

32(4), 341–345 (2005)

59.8 Z. Yan, D. Xu, Y. Li, M. Tan, Z. Zhao: A survey of

the sensing and control techniques for robotic arc

welding, Meas. Control 40(5), 146–150 (2007)

59.9 S.-M.Yang,M.-H.Cho,H.-Y.Lee,T.-D.Cho:Weld

line detection and process control for welding

automation, Meas. Sci. Technol. 18(3), 819–826

(2007)

59.10 U. Dilthey, L. Stein, M. Oster: Through-the-arc

sensing – a universal and multipurpose sensor for

arc welding automation, Int. J. Join. Mater. 8(1),

6–12 (1996)

Part F 59

1040 Part F Industrial Automation

59.11 Y.M. Zhang (ed.): Real-Time Weld Process Monitor-

ing (Woodhead, Cambridge 2008)

59.12 J.P.H. Steele, C. Mnich, C. Debrunner, T. Vincent,

S. Liu: Development of closed-loop control of

robotic welding processes, Ind. Robot 32(4), 350–

355 (2005)

59.13 J.N. Pires, A. Loureiro, G. Bolmsjö: Welding

Robots: Technology, System Issues and Application

(Springer, London 2006)

59.14 R.C. Dorf, S.Y. Nof (eds.): International Encyclo-

pedia of Robotics: Applications and Automation

(Wiley, New York 1988)

59.15 B.-S. Ryuh, G.R. Pennock: Arc welding robot

automation systems. In: Industrial Robotics: Pro-

gramming, Simulation and Applications (Pro

Literatur, Mammendorf 2007) pp. 596–608

59.16 J.N. Pires: Industrial Robots Programming: Build-

ing Applications for the Factories of the Future

(Springer, New York 2007)

59.17 T.J. Tarn, S. Chen, C. Zhou (eds.): Robotic Welding,

Intelligence and Automation, Lecture Notes in Con-

trol and Information Sciences, Vol. 362 (Springer,

Berlin 2007)

59.18 G.E. Cook: Robotic arc welding: research in robotic

feedback control, IEEE Trans. Ind. Electr. 30(3), 252–

268 (2003)

59.19 M. Erickson: Intelligent robotic welding, Tube Pipe

J. 17(3), 34–41 (2006)

59.20 U. Dilthey, L. Stein, C. Berger, K. Million, R. Datta,

H. Zimmermann: Future prospects of shape weld-

ing, Weld. Cutt. 5(3), 164–172 (2006)

Part F 59

1041

Automation i

60. Automation in Food Processing

Darwin G. Caldwell, Steve Davis, René J. Moreno Masey, John O. Gray

Factory-based food production and processing

globally forms one of the largest economic and

employment sectors. Within it, current automa-

tion and engineering practice is highly variable,

ranging from completely manual operations to the

use of the most advanced manufacturing systems.

Yet overall there is a general lag in the use of

automation technology compared with other in-

dustries. There are many reasons for this lack of

uptake and this chapter will initially discuss the

factors that make automation of food production

so essential and at the same time consider coun-

terinﬂuences that have prevented this automation

uptake.

In particular the chapter will focus on the

diversity of an industry covering areas such as

bakery, dairy, confectionary, snacks, meat, poul-

try, seafood, produce, sauce/condiments, frozen,

and refrigerated products, which means that

generic solutions are often (considered by the in-

dustry) difﬁcult or impossible to obtain. However,

it will be shown that there are many features

in the production process that are almost com-

pletely generic, such as labeling, quality/safety

automation, and palletization, and others that do

in fact require an almost unique approach due to

the natural and highly variable features of food

products. In considering these needs, this chapter

has therefore approached the speciﬁc automation

requirements of food production from two per-

spectives. Firstly, it will be shown that in many

cases there are generic automation solutions that

could be valuably used across the industry ranging

from small cottage facilities to large multinational

manufacturers. Examples of generic types of au-

tomation well suited across the industry will be

provided. In addition, for some very speciﬁc dif-

ﬁcult handling operations, customized solutions

will be shown to give opportunities to study the

60.1 The Food Industry............................... 1042

60.2 Generic Considerations in Automation

for Food Processing ............................ 1043

60.2.1 Automation and Safety .............. 1043

60.2.2 Easy-to-Clean Hygienic Design ... 1043

60.2.3 Fast Operational Speed

(High-Speed Pick and Place)....... 1044

60.2.4 Joints and Seals ........................ 1045

60.2.5 Actuators ................................. 1045

60.2.6 Orientation and Positioning ....... 1045

60.2.7 Conveyors................................. 1046

60.3 Packaging, Palletizing,

and Mixed Pallet Automation .............. 1046

60.3.1 Check Weight ............................ 1047

60.3.2 Inspection Systems.................... 1047

60.3.3 Labeling................................... 1048

60.3.4 Palletizing................................ 1048

60.4 Raw Product Handling and Assembly.... 1049

60.4.1 Handling Products That Bruise .... 1050

60.4.2 Handling Fish and Meat ............. 1051

60.4.3 Handling Moist Food Products .... 1052

60.4.4 Handling Sticky Products............ 1053

60.5 Decorative Product Finishing ............... 1054

60.6 Assembly of Food Products –

Making a Sandwich ............................ 1055

60.7 Discrete Event Simulation Example....... 1056

60.8 Totally Integrated Automation............. 1057

60.9 Conclusions........................................ 1058

60.10 Further Reading ................................. 1058

References ..................................................1058

problems/risks/demands associated with food

handling and to provide an insight into the solu-

tion, thereby demonstrating that in most instances

the difﬁcult/impossible can indeed be achieved.

Part F 60

1042 Part F Industrial Automation

60.1 The Food Industry

Food and drink manufacturing forms one of the largest

global industry sectors. In the European Union (EU), it

is in fact the largest manufacturing sector with an an-

nual turnover (in 2006) in excess of € 830billion and

a workforce of 3.8million people [60.1]. However, un-

like in other manufacturing sectors, there is still a very

high level of low-skilled, low-paid labor.

Before considering the detailed useof automation in

food processing, it is important to understand the nature

of the industry. The food industry is not one single sec-

tor making a range of broadly similar products. It is in

fact wide and diverse both in terms of the products and

in structures and is characterized by:

•

A very large number of small and medium enter-

prises (SMEs) operating in a highly competitive

environment

•

Rapid changes in product lines (often several

changes per day)

•

Generally low proﬁt margins

•

Extensive use of manual labor in often unattractive

operating environments

•

Low uptake of automation procedures.

It is also very important to note that, except for a few

multinationals:

•

Engineering research and development activity is

low

•

Ability to exploit and maintain advanced automa-

tion equipment is fairly poor

•

Information technology (IT) and e-Commerce in-

frastructure is generally weak.

Within the industry there is a strong feeling that in

the medium term (3–10years) the number of people

willing to work for the current low wages will decline

and the industry will have to change to survive. Faced

with this problem the industry has identiﬁed the appli-

cation of automation and robotic systems as a major

growth area with the aims of:

•

Improving production efﬁciency and impacting on

yield margins and proﬁtability

•

Reducing waste on all levels: product, energy, pol-

lution, water, etc.

•

Enhancing hygiene standards, and conforming to

existing and future legislation pertaining to food

production, including enhancing hygienic operation

and product traceability

•

Improving working conditions to improve retention

of high-quality, motivated staff

•

Improving the consistency of product quality.

However, despite these very signiﬁcant and com-

pelling driving inﬂuences, only a small number of

companies are yet making signiﬁcant use of automation

for raw and in-progress product handling. The question

therefore arises as to why the food sector should not be

making extensive use of automation. There is no single

simple answer to this question, but the answers seem to

be embedded in a number of technical, ﬁnancial, and

cultural issues [60.2,3].

Although the limited use of automation is cer-

tainly a reﬂection of a conservative investment policy

in a low-margin industry, it is equally clear that in many

instances the use of labor-intensive manual techniques

is a deliberate policy because of:

i) The ﬂexibility provided by the human worker. Hu-

mans handle manipulative complexities with ease,

by combining dextrous handling capabilities (the

human hand), advanced sensing, and behavioral

models of the product accumulated with experience.

ii) A lack ofunderstanding of the properties of thefood

product as an engineering material. The handling

characteristics of many (most) food products can-

not be adequately described with geometry-based

information, as is usually the case with conventional

engineering materials since the geometry of food

product is:

•

In many/most instances nonrigid, often delicate,

and/or perishable

•

Variable in texture, color, shape, and size

•

Often variable as a function of time and the forces

applied

•

Affected by environmental conditions including

temperature, humidity, and pressure

•

Easily bruised and marked when it comes into con-

tact with hard and/or rough surfaces

•

Susceptible to bacterial contamination.

iii) The product deforming signiﬁcantly during han-

dling. Any system developed to handle such food

items therefore needs to react accordinglyto this de-

formation and there is a lack of handling strategies

and end-effectors designed to cope with the variable

characteristics of food products.

Part F 60.1

Automation in Food Processing 60.2 Generic Considerations in Automation for Food Processing 1043

iv) The perceived inabilities of current automation sys-

tems to cope with the variation in product and

production demands.

In addition, the food sector often feels that there are

signiﬁcant issues relating to automation including:

•

Robotic systems and application technology have

been developed for the engineering manufacturing

industry and they cannot transfer across into food

manufacture without signiﬁcant changes.

•

Robot manufacturers and system integrators of-

ten have a poor understanding of the economics,

payback rationale, and operating pressures in the

food industry, which are very different from those

in other sectors that make more use of automa-

tion/robotics. This results in the wrong products

being offered for sale at the wrong price and with

the wrong sales model.

•

Support for complex IT-based systems is largely ab-

sent in smaller food manufacturers and there is no

cost-effective outsourced support available.

•

The space and ﬂexibility requirements of, in par-

ticular, the smaller food manufacturer require that

any automation ﬁts around, andworks with,existing

manual operations. This is in contrast to most en-

gineering automation which is physically separated

from people.

60.2 Generic Considerations in Automation for Food Processing

When considering automation within the sector it is

clear that there are very large differences in the exact

nature of the work and the level and form of automa-

tion. For instance, unlike in the car industry, which

is generally homogenous, making one easily recogniz-

able product, in the food industry this is not true. The

sector can be broken down in many ways, e.g., bak-

ery, dairy, confectionary, snacks, meat, poultry, seafood,

produce, sauce/condiments, frozen, and refrigerated.

Within these areas there are of course many more sub-

divisions and these subdivisions mean that it is almost

impossible to consider the industry as a whole and

certainly from the viewpoint of automation this is ex-

tremely difﬁcult, although there are several key aspects

that have commonality.

60.2.1 Automation and Safety

Issues relating to food safety through accidental or de-

liberate contamination are of paramount concern in all

food manufacturing facilities. To address these concerns

and ensure public conﬁdence there are a number of na-

tional and international standards. Depending on the

country where the food is being manufactured these

standards may be compulsory or voluntary and may be

more or less strictly enforced. Among the most readily

recognized of these standards are

•

Hazard analysis andcritical controlpoints (HACCP)

has been developed as a systematic preventive

approach to food safety that addresses physical,

chemical, and/or biological hazards during the man-

ufacturing process, rather than through end-of-line

and ﬁnished product testing/inspection. This goal is

achieved by identifying potential food safety risks

in the manufacturing process and acting on these

critical control points (CCPs), e.g., by cooking, to

prevent the hazard being realized. HACCP forms

part of the whole manufacturing process including

packaging, and distribution [60.4].

•

Current good manufacturing practice (cGMP) deals

with the control, quality assurance/testing, and

management of food, pharmaceutical, and medical

products in a manufacturing environment [60.5].

•

ISO 22000 aims to bring the structures and bene-

ﬁts of ISO 9000, from which it is derived, to the

food and drink processing and manufacturing sec-

tors [60.6].

The introduction of automation and robotic equip-

ment must of course conform to these standards, ideally

without introducing any new hazards, but at the same

time it is clear that the introduction of automation can

have a positive impact since it permits humans, who

are the most signiﬁcant and certainly the most unpre-

dictable contamination source, to be removed from the

hazard consideration.

60.2.2 Easy-to-Clean Hygienic Design

Machinery to be used for direct handling of food

can be designed following a set of hygienic design

guidelines that ensure good standards of hygiene in pro-

duction [60.7,8], yet there is often a certain degree of

Part F 60.2

1044 Part F Industrial Automation

confusion regarding what constitutes hygienic design

and how hygienic a particular piece of machinery needs

to be. Fundamentally this is product/task speciﬁc but it

is clear that products such as raw meat, ﬁsh, and poultry

are highly susceptible to contamination from microor-

ganisms andrequire veryhigh levels of hygienic design,

while for dryfoods such asbiscuits or cakeslowerlevels

of hygienic design may be more than adequate.

Routine cleaning and disinfection procedures in-

volve the use of acidic, alkaline, and chlorinated

cleaning chemicals. The need for frequent wash-downs

makes a sealed, waterproof structure essential to en-

close and protect internal components. The preferred

material for food processing machinery is 304 or 316

grade stainless steel (BS EN 10088:2005), polished to

a unidirectional satin polish ﬁnish [60.9]. Aluminum

is not sufﬁciently corrosion resistant to commonly

used cleaning chemicals and its use should be avoided

for food-contact applications. Surfaces should be non-

porous and free from cracks, crevices, scratches or

pits that could harbor microorganisms after cleaning.

Painted or coated surfaces should be avoided on food-

contact parts; however, if used, the ﬁnish must be

resistant to ﬂaking or cracking. All parts likely to come

into contact with food should be readily visible for

inspection and accessible for cleaning. Joints that are

screwed or bolted together inherently have crevices that

cannot be adequately cleaned. A rubber seal or gasket

should be used between components joined in this way.

Exposed threads and fasteners suchas screws, bolts, and

rivets should be avoided if possible in food contact ar-

eas. All corners should be radiused and sharp internal

corners should be avoided.

The use of plastics can offer certain advantages over

stainless steel in some applications. However, plastics

are generally more susceptible to failure from a range of

different causes [60.7]. A database of plastics approved

for food use by the US Food and Drug Administra-

tion (FDA) is available online [60.10]. In Europe, the

use of plastics for food-contact applications is regu-

lated by EU Commission Directive 2002/72/EC.One

signiﬁcant disadvantageof plasticsis thatthey are easily

scratched through manual cleaning. Surface scratches

can accumulate over time and harbor microorganisms.

A study byMidelet and Carpentier [60.11]suggests that

microorganisms also attach themselves more strongly

to plastics than to stainless steel. Rubber compounds

used for seals, gaskets, and suction cups should also

be food approved. Nitrile butyl rubber (Buna-N), ﬂu-

oroelastomers (Viton), and silicone rubber are among

those commonly used in the food industry. Likewise lu-

bricants, adhesives, and any other materials that may

come into contact with food should be approved for

food use [60.10,12].

60.2.3 Fast Operational Speed

(High-Speed Pick and Place)

High speed becomes important when an automa-

tion/robotic solution must compete on economic terms

with hard automation or human workers. Low proﬁt

margins in food manufacturing mean that throughput

must be maximized inorder to increase proﬁt. Increased

production capacity also leads to reduced production

costs. Arguably the most common task in food manu-

facturing is pick-and-place handling, where an object

is picked from a conveyor belt and placed into its pri-

mary packaging. The pick-and-place speed of industrial

robots is based on a standard 25×300×25mm

cy-

cle and has been steadily rising. Speeds of between

80 and 120picks/min, which are comparable to that

of a human operator, are now becoming commonplace.

Conveyor belts used in the food industry are generally

no more than 50 or 60cm wide. This width corresponds

to the maximum distance that an operator, standing on

one side of the conveyor, can comfortably reach to pick

an object on the opposite side.

This need for high-speed handling is most acutely

observed in the bakery and confectionary subsectors,

where there is a need for high-speed handling of prod-

ucts that do in general have relatively good structural

formal andrepeatability and can be moved at highspeed

without disintegrating.

Biscuits/cookies form a particularly good example

of this type of product but other breaded products, e.g.,

croissants and even meat precuts such as pepperoni for

a pizza can be considered. In these tasks human op-

erators are required to identify the product (visually),

grasp the product, and place it either into a container or

on to a secondary product. The operating frequencies

are typically high (over 100 picks/min) with motions

in the range of 30–50 cm one way. The mass of the

objects are typically very low (only a few grams) and

when undertaken by humans the user will often pick up

several objects at one instance to minimize the move-

ments to and from the conveyor. Recently the ABB

IRB 340 FlexPicker robothas been extensively and gen-

erally very successfully used in this type of application

to pick up multiple products as a group, or one at a time.

With recent advances in vision technology, robotic

packing lines can handle varying or irregular products

(Fig.60.1). Automation of this type often integrates

Part F 60.2