4 Application Layer Protocols

1
4 Application Layer Protocols

• 4.2 Open Process Control (OPC)

2
OPC Open Process Control
Source http//home.hit.no/hansha/documents/lab/L
ab20Work/opc.htm
3
OPC Open Process Control

• Manufacturer-independent application programming
  interface (API) for automation
• To implement clients which can access plant data
  coming from remote devices (PLCs, field bus
  devices, real-time databases) easily
• Set of commands collected in a software library
  (DLL) to access OPC servers
• OPC clients
• read and write process variables, read alarms and
  events and acknowledge alarms, and retrieve
  historical data from data bases according to
  several criteria
• Implemented on automation platforms (e.g. ABB
  Ax800), which may act themselves as an OPC server
  to publish their data, events and historical
  data.
• OPC server
• Supplied by manufacturer of automation devices
  supplies
• Communicates with its devices through a
  proprietary protocol
• Manages several devices of the same type, several
  servers can run in parallel and each server can
  be accessed by several clients in the same network

4
OPC Overview
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration 4.2.2 OPC Data
Access Objects, Types and properties
Communication model 4.2.3 OPC Alarms and
Events Specification Events Alarm
Conditions 4.2.4 OPC Historical Data
Specification Overview
5
4.2.1 What is OPC ?
OPC (formerly "OLE1 for Process Control", now
"Open Process Control") is an industry standard
set up by the OPC Foundation (http//www.opcfounda
tion.org/) specifying the software interface
(objects, methods) to a server that collects
data produced by field devices and programmable
logic controllers.
APIcovered by the OPC standard
application (OPC client)
servers
OPC serverX
OPC server (simulator)
OPC server Y
Ethernet
Field bus
PLCs Brand X
PLCs Brand Y
Sensors/Actors
1) OLE (Object Linking and Embedding) is a
Microsoft technology for connecting software
components. It has since been extended by the COM
/ DCOM technology. It corresponds to Java Beans.
6
Before OPC
visualization
history data base
Profinet driver
MasterBus MMS driver
XWAY driver
ABB PLCs
Télémécanique PLCs
Siemens PLCs
7
With OPC
Operator
Historian (Information Manager)
application software is written independently
from the type of controller
the drivers still exist, but the clients do
not see them anymore
ABBOPC server
SiemensOPC server
SchneiderOPC server
ProfiNet
MMS
XWAY
ABB AC800M
Télémécanique TSX
Siemens S7
8
Importance of OPC

• Greatest improvement in automation since IEC
  61131.
• More than 150 vendors offer OPC servers to
  connect their PLCs, field bus devices, displays
  and visualization systems.
• Used for data exchange between applications and
  for accessing databases
• DLL for Automation Interface (VB,..) and Custom
  applications (C,..)
• Three major components1) OPC - DA Data Access
  (widespread, mature)2) OPC - AE Alarms and
  Events 3) OPC - HDA Historical Data Access
• and some profiles (batch,) A
  profile is a subset or a specialization of a
  standard to form a stricter standard, adapted to
  an application. E.g. 100 Mbit/s, full-duplex,
  fibre-optical Ethernet is a profile of IEEE 802.3.

9
Specification 1 OPC DA for Data Access
Process variables describe the plant's state,
they are generated by the sensors or calculated
in the programmable logic controllers
(PLCs). Process variables can be sent upon a
change, on demand or when a given time
elapsed. The OPC DA (Data Access) specification
addresses collecting Process Variables.The main
clients of OPC DA are visualization and (soft-)
control.
10
Specification 2 OPC AE for Alarms and Events
Events are changes in the process that need to be
logged, such as "production start" Alarms are
abnormal states in the process that require
attention, such as "low oil pressure"
OPC AE (Alarms and Events) specifies how alarms
and events are subscribed, under which conditions
they are filtered and sent with their associated
messages.The main clients of OPC AE are the
Alarms and Event loggers.
determine the exact time of change (time
stamping) categorize by priorities log for
further use acknowledge alarms(events are not
acknowledged) link to clear text explanation
11
Specification 3 HDA for Historical Data Access
Historical Data are process states and events
such as process variables, operator actions,
recorded alarms,... that are stored as logs in a
long-term storage for later analysis. OPC HDA
(Historical Data Access) specifies how historical
data are retrieved from the logsin the long-term
storage, filtered and aggregated (e.g. compute
averages, peaks). The main client of OPC HDA are
Trend Displays and Historians.
12
Beyond Microsoft OPC UA
In a move to get more independence from Microsoft
and use web technology, a new specification
called " Unified Architecture" (formerly. OPC
XML) that uses web services for all kinds of
transactions query, read, write,
subscribe,... The classical OPC DA, AE and HDA
are implemented with XML / SOAP / WSDL this
allows encryption and authentication of process
data. OPC UA does not only standardize the
interfaces, but also the transmitted data.
13
Client and Servers
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration and
communication 4.2.2 OPC Data Access Objects,
Types and properties Communication
model 4.2.3 OPC Alarms and Events
Specification Events Alarm Conditions Automatio
n Interface 4.2.4 OPC Historical Data
Specification Overview
14
Server(s) and Client(s) in the same node
node
client application (OPC client)
client application (OPC client)
OPC server
OPC server
OPC server
devices
devices
devices
devices
Clients and servers run as parallel
processes The OPC specification defines the
interface between client and server in the
form of objects and methods.
15
OPC for internal communication ABBs SCADA
(800xA) as example
ABB's Operator Workplace (800xA) is at the same
time OPC server and OPC client. Software
components (agents) within AIP expose their
properties as OPC objects. Internal (within the
PC) and external communication (between PCs)
takes place over OPC.
800xA
aspects
aspects
Enterprise Historian
Asset Optimizer
functions
OPC server
Windows PC
process data base
OPC connections
OPC client
ABBOPC server
SiemensOPC server
SchneiderOPC server
16
Direct and Fieldbus access
direct connection
fieldbus connection
client application (OPC client)
client application (OPC client)
(local) OPC server
(local) OPC server
FB Manager
fieldbus
can also be a point-to-point link
proprietary protocol
I/O devices
fieldbus
fieldbus
The OPC server is running all the time, even if
no client is present
FB agent
FB agent
PLC
PLC
17
Accessing a server in another node
client application (OPC client)
stub
DCOM
TCP/IP
DCOM
TCP/IP
TCP/IP
DCOM
DCOM
OPC server
OPC server
Limitation does not work over firewalls. Solution
Tunneller OPC UA
FB Manager
fieldbus
18
Assessment Overview

• What is the objective of OPC ?
• On which technology does OPC rely ?
• What is an OPC Server ?
• What do the main OPC specifications describe ?
• How can an OPC Server access data on another
  machine? And how does it work for OPC clients?

19
OPC Data Access
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration 4.2.2 OPC Data
Access Objects, Types and properties
Communication model 4.2.3 OPC Alarms and
Events Specification Events Alarm
Conditions Automation Interface 4.2.4 OPC
Historical Data Specification Overview
20
OPC DA Item properties

• The process data are represented by three dynamic
  properties of an item
• value numerical or text
• time-stamp time at which this data was
  transmitted from the PLC to the server UTC, not
  local time.
• quality validity of the reading (not readable,
  dubious data, o.k.)
• optional static properties
• description a text string describing the use and
  of the variable (optional)
• engineering unit the unit in which the variable
  is expressed (optional)

Q
T
V
D
U
ID
21
OPC DA Objects as viewed by the OPC server
An OPC server is structured as a directory with
root, branches and leaves (items)
Process Line 1
Tag Name
Controller 1
Level_1
Controller 2
An item is identified by its "fully qualified
ItemID", e.g. "Process_Line_1.Controller_2.Level
_2"
Controller_3.Prog_1
Level_2
Controller_3.Prog_2
Ramp4
Cell 1
Machine 2
Branches may contain other branches and items The
structure may also be flat instead of
hierarchical This structure is defined during
engineering of the attached devices and
sensor/actors. (Intelligent servers can configure
themselves by reading the attached devices)
22
OPC DA Objects as viewed by the OPC client
A client builds its own hierarchy, using the
servers hierarchical view. Items in the server
are defined by the programmer of the PLC A
full-fledged PLC may export some 10000 items, a
client needs only a subset. A client builds
groups, populating them with items it is
interested in. Items of a group are expected to
have similar real-time requirements Groups are
not hierarchical, but flat.
23
OPC DA Mapping items to groups
Each client structures its items by groups,
independently from the server. Initially, the
client browses the server structure to check if
the items it is interested in exist. A client
registers its groups and items at the server.
The server keeps the structure of all its
clients.
clients
Client1
Client2
GroupX
GroupZ
Item1
Item2
Item3
Item1
Item2
Server root
server
Area 1
Temperature
Oven_1
Heat_On
Level
Tank_1
Empty_Valve
Area 2
Fill_Valve
Area 51
24
OPC DA Communication Model
4.2.1 OPC Common Overview usage and
specifications Clients and Servers
configuration 4.2.2 OPC Data Access Objects,
Types and properties Communication
model 4.2.3 OPC Alarms and Events
Specification Events Alarm Conditions Automatio
n Interface 4.2.4 OPC Historical Data
Specification Overview
25
4.2.2 OPC DA Example of access to a variable
controller program
OPC application
Reactor_1.Program2
ReadItem ("OPCReactor1 Program2.MotorSpeed")
MW1003
MotorSpeed
MW1004
Temperature
Value 112

.
symbols
load symbol table
OPC server
code
Get () 192.162.0.2, MW1003)
Return (MW1003, 112)
Network
Program 2
Reactor_1
Marker MW1003
26
OPC DA Read Communication Models (per group)
synchronous
asynchronous
client
server
server
client
Call
Call
myGroup.SynchRead()
myGroup.AsyncRead()
Reply
Reply
myGroup_AsyncReadComplete()
server
client
Subscribe
myGroup.IsSubscribed
on change ("subscription-based")
Notify
myGroup_DataChange()
Notify
myGroup_DataChange()
27
OPC DA Write Communication Models (per group)
client
server
server
client
Call
Call
myGroup.SynchWrite()
myGroup.AsyncWrite()
Reply
Reply
myGroup_AsyncWriteComplete()
The OPC interface accesses only groups, not
individual items.
28
OPC DA communication paradigm
OPC DA works according to the shared memory
paradigm. This means that a newer value
overwrites the older one, no queues or history
are kept. The server does not guarantee that
different clients see the same snapshot of the
plant. The server does not guarantee that all
changes to variables are registered, changes may
be missed if the polling period is too long.
OPC DA Client
OPC DA Client
OPC DA Server
29
OPC DA Assessment

• How does the OPC server know a) where to fetch an
  item? b) which items belong to which group?
• What are the DA the read and write operations ?
• Is communication done by items, by groups or by
  collection of groups ? Why?
• Can a change of an OPC variable be notified as an
  event, or shall the client poll ?
• What are the implications of the shared memory
  paradigm for the application developer?

30
OPC Alarms and Events
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration 4.2.2 OPC Data
Access Objects, Types and properties
Communication model 4.2.3 OPC Alarms and
Events Specification Events Alarm
Conditions 4.2.4 OPC Historical Data
Specification Overview
31
Alarms and Events Purpose
The controllers (PLC) generate events in response
to changes in the plant variables. together with
their precise time of occurrence, type, severity
and associated message for the human operator.
An OPC AE server registers these events and
makes them available to several clients Alarms
are described as state machines and may require
acknowledgement. The OPC Alarms Events
Interface gives access to the AE server, allowing
to - browse the OPC AE Server for predefined
events. - enable or disable alarms and events -
subscribe to alarms and events of interest -
receive the event and alarm notifications with
the associated attributes - acknowledge alarms
32
AE Definitions
An event is a general change of state that is
relevant to the OPC server. An event signal a
change 1) in the field device ("production
started") 2) in the OPC server ("alarm
acknowledged") 3) in the application ("operator
action") An alarm indicates a state of the
process that requires attention and is relevant
to the OPC server. An alarm is represented by an
alarm condition, which is a state machine
determining if the alarm has been enabled,
triggered or acknowledged. An alarm rises several
events. An event or an alarm does not transmit
process values, but boolean information
indicating a change of state, its originator, the
time of its occurrence and a message intended for
a human operator. Alarms and events may not get
lost (contrarily to OPC DA, which does not
guarantee completeness) Alarms and event are
precisely time-stamped by their source,
(contrarily to process variables, which are
time-stamped by the receiving OPC server).
33
AE communication paradigm
OPC AE works according to the message passing
paradigm, contrarily to OPC DA, that works
according to the "shared memory" paradigm. This
means that an event is kept in a queue until all
clients have read it (or timed out). The AE
server guarantees that different clients will see
all events in the same sequence.
OPC AE Client
OPC AE Client
OPC AE Server
1234 23.114
1234 32.334
34
AE Displaying Alarms and Events
Alarms and events are usually displayed
differently on an operator screen. - Events are
displayed in an event list that can become quite
long (typically 1000 entries), entries are not
cleared when the source of the event returns to
normal - Alarms are displayed in a short list
(typically 50 alarms) appearance changes when
the alarm is acknowledged, an alarm line is
cleared when the alarm signal is cleared (but
remains in the log).
Ackcheckbox
35
AE Events
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration 4.2.2 OPC Data
Access Objects, Types and properties
Communication model 4.2.3 OPC Alarms and
Events Specification Events Alarm
Conditions 4.2.4 OPC Historical Data
Specification Overview
36
AE Events kinds

• OPC AE defines three kinds of events
• simple process control system related events
  (change of a boolean variable)
• condition-related notifies a change of an alarm
  condition (CLEARED, ACKNOWLEDGED), (see later)
• tracking-related origin outside of the process
  (e.g. operator intervention)

37
AE Event- identification
An event is identified by - its source (the
object that generates the event. e.g. Tank1) and
- the event name (which can be the same as in
another object, e.g. HiLevelCond)
HiLevelCond
HiLevelCond
Function Block
event
event
LoLevelCond
LoLevelCond
event
event
event name
Tank1
Tank2
event signal (boolean expression) is an external
signal to be used ?(boolean) signal name for
external signal (20 characters) name of the
source (30 characters) message (60 characters)
38
AE Events - Notification
Tank1LevelHigh_SimpleEvent(source, timestamp,
message, severity, category)
AE Client
specified communication COM/DCOM
OPC AE Server
queue
unspecified communicationnetwork, fieldbus or
internal bus
event notification
message
Controller
timestamp
EventFB
Tank1
Plant
Level Switch
39
AE Events - Time Stamp
There are three places where events can be
time-stamped - at the device that originally
produced the data (external event - low-level
event) allowing Sequence-Of-Events with a high
accuracy, down to microseconds - at the
controller, (internal event) using the
controller's clock to time-stamp messages
giving accuracy not greater than the period of
the tasks, about 1 ms. - at the OPC Server, when
an event message arrives (tracking events) not
more accurate than DA, about 10 ms)
The OPC server can be configured to register the
time stamp at the instant of the event transition
(positive or negative) and the instant of the
acknowledgement.
40
AE Alarm conditions
4.2.1 OPC Common Overview usage and
specifications Clients and Servers
configuration 4.2.2 OPC Data Access Objects,
Types and properties Communication
model 4.2.3 OPC Alarms and Events
Specification Events Alarm Conditions 4.2.4
OPC Historical Data Specification Overview
41
AE Alarms - Condition Definition

• An (alarm) condition is described in a named
  state machine
• The condition state is defined by three
  variables
• Enabled the condition is allowed to send event
  notifications
• Active the alarm signal is true
• Acknowledged the alarm has been acknowledged

Alarm signal(e.g. FIC101.PV gt 100 AND FIC101.PV
lt 150)
Condition
Acknowledgement signal(a positive transition of
a boolean variable)
Condition state
Enable (positive transition)Disable (positive
transition)
42
AE Alarms - Condition states and acknowledgement
event notification
alarm signal
acknowledgement
InactiveAcked
ActiveUnacked
ActiveAcked
InactiveAcked
ActiveUnacked
InactiveUnacked
InactiveAcked
InactiveUnacked
condition state
alarm_signal ?
EnabledInactiveAcked
condition state transitions (here always
enabled)
alarm_signal ?
EnabledActiveUnacked
EnabledActiveAcked
Ack ?
Ack ?
alarm_signal ?
EnabledInactiveUnacked
alarm_signal ?
An event is generated each time the alarm signal
changes state, or is acknowledged
43
AE Alarms - Acknowledgement
An alarm condition becomes active when the PLC
produces an alarm signal describing an abnormal
state defined by the application (e.g. the level
of the tank is too high). The operator is
expected to acknowledge this condition (client
ack)Alternatively, a local operator may use a
button or HMI that the PLC reads (field ack)
event notification
AE Client
Tank1Level_ConditionEvent
client ack (acknowledger ID)
COM / DCOM
OPC AE Server
Network, field bus,or internal bus
message
Condition
time-stamp
LevelHigh
controller
AckButton(field ack)
Alarm Signal
44
AE Summary alarms and events
Event
Alarm
AE Client
AE Client
COM / DCOM
OPC AE Server
OPC AE Server
event notification (source, timestamp, message)
alarm notification (source, timestamp, message,
condition, subcondition, severity, type)
message
Condition
timestamp
message
EventFB
controller
controller
ack
alarm
event
45
OPC AE Assessment

• What is the difference between Alarms and Events?
• Where are Alarms and Events time stamped?
• How does the message passing paradigm influence
  the OPC client application developer?

46
OPC Common Overview
4.2.1 OPC Overview Usage and specifications Clie
nts and Servers configuration 4.2.2 OPC Data
Access Objects, Types and properties
Communication model 4.2.3 OPC Alarms and
Events Specification Events Alarm
Conditions 4.2.4 OPC Historical Data
Specification Overview
47
Historian Example
GE Fanuc/Intellution iHistorian (iFix...)
Questions to the historian What was the value of
FIC101 last week ? What was the flow average
during October ? Which were the daily averages in
October ? What was the total flow in each month
? How much fuel did we use for the batch ? Give
the answers in form of tables, pie diagrams,
spreadsheet, reports
Features Unlimited Point CollectionSub-Second
Data Collection Rates Enhanced Data
CompressionTrue Thin Client AdministrationFault
Tolerant Architecture
Data collection, archiving and retrieval Report
generation Computations (e.g. VBScript) Secure
access (FDA 21 CFR 11) 20'000 actions/ s, Up to
100'000 data points
48
HDA Historical Data Access
e.g. Event Logger
e.g. Trend Analysis
HDA Clients
independent processes
OPC HDA Server
history database
hidden
calculations
raw and ordered data
collector
proprietarydata acquisition
OPC DA Server
Field device
Field device
49
HDA Purpose
An OPC HDA server gives access to a historical
data base (logs) in which data from the process
have been collected and time-stamped, possibly
through an OPC DA interface. The OPC HDA
interface clients, such as trend analysis,
product tracking or data mining, that require
ordered access these data logs. The OPC HDA
interface allows to - browse the historical
data base - retrieve data through proper
filtering, e.g. by date range, by identity, by
property - build aggregates over the retrieved
data, such as average, minimum, maximum. -
enter new entries, correct entries or remove
entries - enter / delete annotations in the
history data base
50
HDA Raw log
12.3.02 1340
Gpcpt2ofpbonne
4824
By definition, values are registered when they
change.(even if data are acquired by periodic
polling)
12.3.02 1340
Cpt2bac
50
12.3.02 1340
Gpcpt2bac
70
12.3.02 1340
Gpcptbe2
45
12.3.02 1341
Gpcpt1bac
151
12.3.02 1341
Gpcpt1ofpbonne
4826

• Data in the historical database are identified by
  their
• itemID (here, represented by their name),
• value, (of the respective type)
• quality (good, stale, bad), and
• timestamp (UTC).

12.3.02 1341
Gpcptae2
45
12.3.02 1341
Cpt1bac
49
12.3.02 1341
Gpdefr2
64
12.3.02 1341
Gpvoydef
2
12.3.02 1341
Gpr3tempscycleprd
318
12.3.02 1342
Gpstn1e1
16
12.3.02 1342
Gpalarme1
0
12.3.02 1342
Gpalarme2
0
12.3.02 1343
Gpetatmodemarche
2
12.3.02 1343
Gptpscycle
1346
12.3.02 1343
Gpetatmodemarche
1
12.3.02 1343
Gpdefgene1
16
12.3.02 1343
Gpetatmodemarche
0
12.3.02 1343
Gptpscycle
317
12.3.02 1343
Gpdefr2
0
12.3.02 1343
Gpvoydef
0
12.3.02 1343
Gpdefgene1
0
12.3.02 1344
Gpetatmodemarche
1
12.3.02 1344
Gpr2tempscycleprd
1992
12.3.02 1344
Gptpscycle
435
12.3.02 1344
Gpalarme3
1
12.3.02 1344
Gpalarme4
1
12.3.02 1344
Gpalarme3
0
12.3.02 1344
Gpcpt2ofpbonne
4823
51
HDA How to reduce raw data (even before HDA
comes into play)

• Data sent to the OPC DA server on change,
  collector records data in a circular buffer log
• Since storage capacity is limited, data are
  reduced by
• if a variable is received more often than the
  log's minimal storage interval (e.g. 1s), the log
  keeps the latest of all values of the interval
• if (and only if) a received variable changed by
  more than the "exception deviation" (e.g. 5,
  analog values only), it is entered into the log.
• if a variable changed by less than the exception
  deviation, it may be forced into the log after
  the log's maximum storage interval (e.g. 4 s)
  elapsed.

min time
max time
max time
max time
max
exception deviation
min
time
process value (event)
log entry
If the time scale of the log is smaller than that
of the trend display in this case, values have to
be interpolated to be displayed correctly
52
HDA Application Trend Display

• Parameters
• time scale (with possible offset, zoom, pan)
• amplitude scale (low range, high range, scale
  units)
• style smoothed, stepped, filled (several ways
  to display the same data)
• extrapolate how to display values not received
  (e.g. because they did not yet change)

log how were data sampled
53
HDA Hierarchical logs
long-term log 1 hour_forever ? 5.2 MB / Year
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
1
2
3
27
28
29
30
octobre
october
september
1 mn_7 days log ? 6 MB
1s_24 hours log 50 traces _at_ 12 B ? 52 MB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
A hierarchical log is built on the data contained
in the parent log. To reduce the log size,
several aggregations can be applied - record
only maximum, minimum, average over a period,
etc...
Actual data
54
Toy Examples Logging Temperature
http//home.hit.no/hansha/documents/subjects/IA62
09/project/IA620920Project20Work20-20Weather2
0System.pdf
55
Real-Life Examples and Exercise
Integration of redundant GE Mark V turbine
controllers at a power plant
The controllers were going to tie into the
plants main DCS, an ABB Advant control system.
Process Portal B and MicroSCADA client
applications were used to visualize process data
from the turbine controllers. Exercise Draw
schema of system and describe which OPC
specifications are used and how.
Wind Generation
AES Wind Generation manages 7 different wind
farms across the United States, five of which
comprise more than 500 turbines, with a total
generating capacity of over 700 MW. There are six
different turbine models, from four different
manufacturers. Exercise Explain which OPC
concepts are useful in this context and how they
can be applied to support a SCADA system
http//www.matrikonopc.com/portal/downloads.aspx?d
ID132 and https//www.matrikonopc.com/portal/down
loads/case_studies/AESWindGeneration_OPC.pdf
56
To probe further.
OPC Foundation Specifications
http//www.opcfoundation.org SoftwareToolbox
Examples in Visual Basic http//www.softwareto
olbox.com/Tech_Support/TechExpertiseCenter/OPC/opc
.html The Code Project OPC and .NET
http//www.codeproject.com/useritems/opcdotnet.asp
Matrikon Free client and server
http//www.matrikon.com WinTech Toolkit for
an OPC server http//www.win-tech.com/html/opc
stk.htm NewAge Automation Toolkit for an OPC
server http//www.newageautomation.com