1. 4 Application Layer Protocols
EPFL, Spring 2017
4 Application Layer Protocols
Last week, field bus principles and operation, this time what to use fieldbusses for and how…

2. Device Management Protocols
Application layer protocols offering remote services for large networks of devices:
Monitoring (health of devices and network, get current state)
Management (deployment, configuration, change settings)
current loop mA
HART
SNMP: Simple Network Management Protocol
MMS: Manufacturing Messaging Specification

3. The classical solution for analogue values
Current Loop
The classical solution for analogue values
Example differential pressure transducer
4..20 mA current loop
Reminder:
This is the simplest way of reading a variable.
Only one variable can be read, but by multiple readers.
If there is a circuit problem the current is 0 and can be detected (but not identified where and why).
fluid

4. 4.1.2 HART
Data over mA loops
Practically all 4..20mA devices come equipped with HART today
About 15 Mio devices are installed worldwide.
Reminder:
Reminder:
Next step up from 4-20mA. As discussed in more detail last time, this protocol enables basic master-slave communication.
more info:

5. The Round card Standardized printed circuit board
Standardized printed circuit board
can be mounted in instrument
contains modem, processor, RAM, EPROM… necessary to execute the HART protocol.
Why round? Most hydraulic instruments have a round case.
One of the reasons for its success is the round card, a standardized printed circuit board.
It can be mounted on almost any instrument, containing the modem, a processor, RAM, EPROM and all the logic and software necessary to execute the HART protocol.
It is round because most hydraulic instruments have a round case.

6. HART commands summary Universal Commands Common Practice Commands
Device-Specific Commands (example)
• Read manufacturer and device type
• Read variable and units
• Read current output and percent of range
• Read up to four predefined dynamic variables
• Read/write tag, descriptor, date
• Read/write 32-character message
• Read device range values, units, and damping time constant
• Read or write final assembly number
• Write polling address
• Read selection of up to four dynamic variables
• Write damping time constant
• Write device range values
• Calibrate (set zero, set span)
• Set fixed output current
• Perform self-test
• Perform master reset
• Trim variable zero
• Write variable unit
• Trim DAC zero and gain
• Write transfer function (square root/linear)
• Write sensor serial number
• Read or write dynamic variable assignments
• Read or write low-flow cut-off
• Start, stop, or clear totalizer
• Read or write density calibration factor
• Choose variable (mass, flow, or density)
• Read or write materials or construction information
• Trim sensor calibration
• enable PID, write PID setpoint
• Valve characterization
• Valve setpoint
• Travel limits
• User units
• Local display information
Number of commands limited, most of them focus on reading,
some prompt the device to change settings or perform a self-test.

7. Device access hand-held field device SCADA network adapter
type
manufacturer
network adapter
device
volumetric flow rate
FlowPro
ABB
velocity
13.32
m2/s
diff. pressure
9.8 Pa
density
0.8
kg/l
cross sectional area:
pipe inside diameter
3 cm2
2 cm
hand-held
field device
SCADA
network adaper
network
adapter
modem
Apart from the obvious master, the SCADA system, HART handheld devices are often used to check
values and parameters in the field.
4-20 mA loop for HART
13.32
9.8
0.8

8. SNMP:Simple Network Management Protocol
IETF (Internet standard) protocol for device and network management (routers, switches, servers, workstations, printers…).
prevent down-time
discover bottlenecks
identify conditions that need an operator’s attention.

9. Simple Network Management Protocol
IETF (Internet standard) protocol for device and network management (widely supported, especially by routers, switches, servers, workstations, printers…).
Configuration Management
Keeping track of device settings
Fault management
Dealing with problems and emergencies (router stops forwarding, server loses power, etc)
Performance Management
How smoothly is network running?
Can it handle the current workload?
This includes configuration management (to keep track of device settings),
fault management (dealing with problems and emergencies, e.g., when router stops forwarding, server loses power, etc) and
performance management (to quantify how smoothly a network is running and how well it handles the current workload).

10. SNMP - MIB objects networked device MIB
Device contains MIB (managed information base) and an agent to access MIB
Agent
TCP/UDP/IP (171 objects)
NT network (90 objects)
DHCP (14 objects)
WINS (70 objects)
Appletalk
Nowell
IPX
DecNet
…..
CISCO (proprietary)
Each networked device contains a MIB (managed information base).
In this MIB there are objects that describe
parameters,
statistics and
state of device
(mostly communication related).
Mostly parameters, statistics and error counters used for communication

11. SNMP – ASN.1 Object example
tcpMaxConn OBJECT-TYP
SYNTAX Integer32 (-1 | )
MAX-ACCESS read-only
STATUS current
DESCRIPTION
"The limit on the total number of TCP connections the entity
can support. In entities where the maximum number of
connections is dynamic, this object should contain the
value -1."
::= { tcp 4 }
tcpActiveOpens OBJECT-TYPE
SYNTAX Counter32
"The number of times that TCP connections have made a direct
transition to the SYN-SENT state from the CLOSED state.
Discontinuities in the value of this counter are
indicated via discontinuities in the value of sysUpTime."
::= { tcp 5 }

12. Example SNPM Network

13. SNMP - Access to Managed Objects
User
object interface
MIB
managed
information
base
Manager
Agent
call
(indication)
(request)
reply
(confirm)
(response)
UDP
UDP IP IP
ISO Type 1
ISO Type 1
ISO (Ethernet)
ISO (Ethernet)
management
messages
network

14. SNMP - Operations on objects
Operations (PDU type):
Get (read)
Set (write)
GetNext (transversal reading)
GetBulk (optimized GetNext, v2 and v3)
Response (variable bindings and acknowledgement)
Trap (asynchronous agent notification, priorities)
Since SNMPv1/SNMPv2 do not provide authentication, “Set” commands are normally disabled. Traps are rarely used.

15. Example Network and Queries

16. SNMP - How are objects identified ?
ISO defined a world-wide addressing scheme on a hierarchical basis:
MIB objects are identified by a concatenation of numerical identifiers
quite wasteful, but bearable in LANs

17. SNMP example of identification
== .iso.org.ieee.standards-association-c-series-standards.std-c37.part238.
ieeeC37238TSMib.ieeeC37238Objects.ieeeC37238DefaultDS.ieeeC37238DefaultDSClkIdentity

18. SNMP - Assumptions about the underlying communication network
- the network is connectionless (datagrams): only UDP is used (no TCP).
- manager and agent can send messages to each other spontaneously
- all entities must be able to receive and send packets of at least 484 octets
- the network supports broadcast
Further reading:

19. Manufacturing Messaging Specification (MMS)
schedule
Program
Invocation
Named
Variable
Variable List
File
Types
Semaphore
Event
Enrolment
Transaction
Domain
Operator
Station
Journal
Condition
robot configuration
Action
Messaging system for transfer of
process data and supervisory control
information between networked devices and/or computer applications.
MMS is very popular in manufacturing, especially in large plants
with numerous conveyor belts, robots, CNC machines, paint shops, logistics.
devices download schedules and tasks from production management, with a connection to administration.

20. Interaction between Operator Workplace and field equipment
SCADA (client) (any technology)
network (any)
controller (server)
(any technology)
represents automation objects, i.e. a collection of PLC1 variables
manufacturing devices
represent pieces of equipment
Info from field to SCADA, by defining objects which must exist in every conformant device,
on which operations can be executed (read write)
start a transmission spontaneously
MMS: access all controllers, regardless of the manufacturer, in the same way.

21. The basic MMS idea: read a variable
client (any technology)
request
read
variable name
response
value
status
network (any)
server
MMS defines
Objects
Messages
Encoding rules
Protocol
I / O devices
basic MMS idea: read and write equipment variables using standard messages.

22. MMS - Manufacturing Message Specification
Developed 1980 for General Motor’s flexible manufacturing initiative
Reputation: heavy, complicated and costly (due to poor implementation)
But:
Boeing adopted MMS (on Ethernet)
Adopted by the automobile industry and power distribution
Standardized as:
[1] ISO/IEC : Industrial Automation systems – Manufacturing Message Specification -
Part 1: Service Definition (IS 1990)
[2] ISO/IEC : Industrial Automation systems – Manufacturing Message Specification -
Part 2: Protocol Specification (IS 1990)
messaging system for transferring process data and supervisory control information between networked devices and/or computer applications. 

23. MMS - Manufacturing Message Specification
Application
MMS does not specify the application interface
MMS specifies a set
of messages which
to control an MMS
server
allow an MMS client
device
device
MMS specifies a set
of objects an MMS
server contains
MMS
MMS specifies how messages are encoded for transmission
MMS
client
server
request
(command)
response
(reply)
communication
stack
communication
stack
• A set of standard objects which must exist in every conformant device, on which operations can be executed (example: local variables, read and write) or which can start a transmission spontaneously
• A set of standard messages exchanged between a manager and an agent station for the purpose of controlling these objects
• A set of encoding rules for these messages
• A set of rules for exchanging messages between devices (basic protocol)
MMS does not specify application-specific operations (e.g. change motor speed). This is covered by application-specific, “companion standards” (e.g. flexible manufacturing, drives, remote meter reading)
network
linking
device
router

24. MMS - Communication Stack
Association Control Service Element, ACSE, ISO 8649/8650, N2526,N2327
Application
Abstract syntax notation,ISO 8822/8823, 8824/8825
Presentation
ISO 8326/8327
Session
ISO 8073 Class 4
Transport
ISO 8473 connectionless
Network
ISO Type 1
Link
ISO (Ethernet)
ISO
MAC
(token bus)
Physical
quite heavy… Boeing decided to drop ISO for TCP/IP ("TOP“).

25. MMS in the fieldbus stack context
time-critical
applications
(PLC tasks)
time-benign
applications
(HMI, download)
real-time
data base
time-benign
messages
time-critical
process variables
MMS 7
Application (Association)
implicit
implicit (ASN.1) 6
Presentation
real-time
traffic (not MMS)
Remote Procedure Call 5
Session
connection-oriented 4
Transport (connection-oriented)
connectionless 3
Network (connectionless)
connectionless
2"
Logical Link Control
medium access 2'
Link (Medium Access)
common
media 1
Physical
MMS is not for real-time communication, but it can access the real-time variables

26. MMS - Basic Communication Principles
MMS assumes that the communication stack offers two services:
Indication
MMS Responder
(server)
Request
Confirmation
Response
network
processing
event
MMS Requester
(client) 1)
Remote Procedure Call
(Call paired with reply,
synchronous, unicast) 2)
Event Reporting
(spontaneous messages sent
by server)
Addressing not specified, messages contain communication reference to identify connection. Usually, clients/servers are addressed by IP address and MMS server uses port number 102.

27. MMS - Event services MMS provides services to:
- Event Condition (define boolean condition that triggers event and its priority)
- Event Enrolment (define MMS client(s) to notify when event is triggered)
- Event Action (define MMS service to be executed when event occurs)
MMS client
MMS client
DefineEventCondition
EventNotification
AlterEventCondition
AckEventNotification
enables/disables event conditions
(MMS server)
Event
Enrolment
Event
Condition
Event
Action
Who?
When?
What?
event notification and confirmation
Events are the most complicated part of MMS

28. MMS - Event triggering
MMS client
MMS client
TriggerEvent
EventNotification
AckEventNotification
MMS Server
NETWORK-TRIGGERED
Event
Enrolment
Event
Condition
Event
Action
boolean variable
MONITORED
cyclic monitoring
plant
Events are triggered by a change in a boolean variable in the server (monitored event) or by MMS client (trigger event) as an invitation procedure

29. VMD: Virtual Manufacturing Device
Definition of objects, services, and behavior
Only specifies the network-visible aspects (device / application communication)
Internal implementation details (programming language, operating system, CPU type, input/output (I/O) systems) not specified by MMS
interoperability
Application Programming Interface
(MMSI = MMS interface)
Virtual
Device
communication
stack
physical
link
network
transport
session
presentation
robot
flow meter
cell
For increased interoperability MMS offers
A virtual device represents a (complex) piece of equipment
With application-specific object models
A physical device (PLC) may implement one or more virtual devices

30. MMS - Importance
Used as reference model for industry, not just implementation.
High complexity makes it very general, but required bandwidth and computing power were out of reach until few years ago.
Part of almost every PLC today.
Gave rise to several other "simpler" models (DLMS, BacNet, FMS....)
Base of IEC „Communication networks and systems in substations“ together with TCP/IP/Ethernet.
MMS has been a reference model for industry rather than simply an implementation.
Its high complexity makes it very general, but the requested bandwidth and computing power were out of reach until few years ago.
It is part of almost every PLC today.
It gave rise to several other "simpler" models (DLMS, BacNet, FMS....)
It is the base of IEC „Communication networks and systems in substations“, which bases on TCP/IP/Ethernet.
For more information, see:

31. Exercise Fill in and extend the table Criteria Real time Monitoring
Management
Event handling
Comm. Pattern
Security
Application area …
Criteria
4-20mA loop
HART
SNMP
MMS
Bandwidth, reliability, cost

32. Assessment What is the purpose of the HART protocol ?
Which communication is used between a hand-help and a field device ?
Which categories of HART commands do exist ?
What distinguishes Hart from SNMP ?
What distinguishes SNMP from MMS ?
What are the (dis)advantages of MMS ?

33. Open Process Control (OPC)
(formerly: "OLE1 for Process Control", now: "Open Process Control")
• Learn the advantages of using OPC
• Understand the OPC Client/ Server architecture
• Review the milestones in the history of OPC
• Understand how to create a connection between an
OPC Client and OPC Server
• Recognize what factors can affect OPC communication
• Review the most widely implemented OPC
specification: OPC DA

34. Open Process Control: API for Automation
Manufacturer-independent application programming interface (API)
To implement clients which can access plant data coming from remote devices (PLCs, field bus devices, real-time databases) easily
Set of commands to access OPC servers
OPC clients
Read and write process variables, read alarms and events, acknowledge alarms, retrieve historical data from data bases according to several criteria
Implemented on automation platforms (e.g. ABB Ax800), which may act themselves as OPC servers 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, each server can be accessed by several clients in same network
Interoperability, different vendors => API for autoamtion

35. OPC Specifications 4.2.1 OPC Overview Usage and specifications
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. 4.2.1 What is OPC ?
Industry standard set up by the OPC Foundation ( specifying the software interface (objects, methods) to a server that collects data produced by field devices and programmable logic controllers.
API covered by the OPC standard
application
(OPC client)
servers
OPC server X
OPC server
(simulator)
OPC server Y
Ethernet :
Field bus
Manufacturer-independent application programming interface (API
- implement clients accessing 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 A&E, and retrieve historical data
Implemented on automation platforms (e.g. ABB Ax800
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
PLCs Brand X
PLCs Brand Y
Sensors/Actors

37. Before OPC visualization history data base ABB PLCs Télémécanique PLCs
MasterBus
MMS driver
Profinet
driver
XWAY
driver
ABB PLCs
Télémécanique PLCs
Siemens PLCs

38. (Information Manager)
With OPC
Operator
Historian
(Information Manager)
application software is written independently from the type of controller
Proprietary drivers still exist, clients do not see them anymore
ABB OPC server
Schneider OPC server
Siemens OPC server
MMS
XWAY
ProfiNet
ABB AC800M
Télémécanique TSX
Siemens S7

39. 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
Three major components: 1) 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

40. Specification 1: OPC DA for Data Access
Process variables describe a plant's state, generated by the sensors or calculated in the programmable logic controllers (PLCs).
Process variables sent upon a change, on demand or when given time elapsed. OPC DA (Data Access) specification addresses collecting Process Variables. It provides services for reading and writing Process Variables. Main clients of OPC DA are visualization and (soft-) control.

41. Specification 2: OPC AE for Alarms and Events
Events are changes that need to be logged, such as "production start"
Alarms are abnormal states that require attention, such as "low oil pressure"
OPC AE (Alarms and Events) offers services to specify alarms and events, their subscription, 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

42. Specification 3: HDA for Historical Data Access
Historical Data are process states and events (process variables, operator actions, recorded alarms,...) stored as logs in a long-term storage for later analysis.
OPC HDA (Historical Data Access) offers services to specify how historical data are retrieved, filtered and aggregated (e.g. compute averages, peaks). The main client of OPC HDA are Trend Displays and Historians.

43. Server(s) and Client(s) in the same 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.

44. OPC for internal and external communication
ABB SCADA 800xA as an example
Enterprise
Historian
Asset Optimizer
External communication
800xA
Internal communication
aspects
aspects
functions
OPC server
process
data base
ABB OPC server
Schneider OPC server
Siemens OPC server
OPC client
External communication

45. OPC UA (IEC 62541)
OPC UA is a service-oriented communication framework for the exchange of information models, replacing OPC “Classic” which was based on Microsoft DCOM communication.
Service provider receives requests, processes them and sends results back.
In contrast to classic Web services that describe their services over a WSDL and can thus be different with each service provider, generic services are already defined with OPC-UA

46. OPC UA Data Model
Specialized models extend the base model by adding functionality like DA, AE, HDA,...

47. OPC UA Communication
OPC UA does not only standardize the interfaces, but also the transmitted data and enables encryption and authentication of process data.
Address space is structured hierarchically for interoperability. Top levels are standardized for servers.
Two main protocol bindings available:
optimized TCP-based binary protocol
HTTP/HTTPS web service with binary or XML messages
Discovery mechanisms:
identification of devices and their functions within a network.
aggregation across subnets and intelligent, configuration-less procedure (e.g. Zeroconf) to identify and address network participants.
More info:

48. Assessment Overview What is the objective of OPC ?
What is an OPC Server , what an OPC client ?
What do the main OPC specifications describe ?
On which technologies does OPC UA rely ?

49. OPC Data Access 4.2.1 OPC 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
Automation Interface
4.2.4 OPC Historical Data Specification
Overview

50. OPC DA: Objects as viewed by the OPC server
root
branch
leaf (item)
Level_1
Level_2
Ramp4
TAG
fully qualified ItemID Process_Line_1.Controller_2.Level_2
Process Line 1
Controller 1
Controller 2
Controller_3.Prog_1
Controller_3.Prog_2
Cell 1
Machine 2
- An OPC server is structured as a directory with root, branches and leaves (items)
Branches can contain other branches (flat possible)
Intelligent servers can configure themselves by reading the attached devices.
- An item is identified by its fully qualified ItemID, e.g. "Process_Line_1.Controller_2.Level_2"
Structure is defined during engineering of the attached devices and sensor/actors.

51. OPC DA: Objects as viewed by the OPC client
full-fledged PLC may export some 10’000 items
client needs only a subset.
client builds flat groups of items with similar time requirements
full-fledged PLC may export some 10’000 items
client needs only a subset.
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.
GroupX
GroupZ
...
Item1
Item2
Item3
Item1
Item2

52. OPC DA: Mapping items to groups
clients
client registers groups at server.
server keeps structure of all its clients.
Client1
Client2
GroupX
GroupZ
Item1
Item2
Item3
Item1
Item2
Server root
server
Area 1
Oven_1
TAG
Temperature
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.
TAG
Heat_On
Tank_1
TAG
Level
Area 2
TAG
Empty_Valve
TAG
Fill_Valve
Area 51

53. 4.2.2 OPC DA: Example of access to a variable
OPC application
controller program
Reactor_1.Program2
MW%1003
MotorSpeed
MW%1004
Temperature … ….
load
symbol
table
symbols
ReadItem ("OPC:Reactor1: Program2.MotorSpeed")
OPC
server
code
Get () , MW%1003)
Return (MW%1003, 112)
Network
Program 2
The OPC server loads the symbol table of deployed programs.
When an OPC client app wants to access a process variable,
the OPC server can map it to a call understandable by the PLC.
Reactor_1
Marker: MW%1003

54. OPC DA: Read Communication Models (per group)
synchronous
asynchronous
client
server
client
server
Call
Call
myGroup.SynchRead()
myGroup.AsyncRead()
Reply
Reply
myGroup_AsyncReadComplete()
client
server
Subscribe
myGroup.IsSubscribed
on change
("subscription-based")
The OPC interface accesses only groups, not individual items.
Notify
myGroup_DataChange()
Notify
myGroup_DataChange()

55. OPC DA: Write Communication Models (per group)
client
server
client
server
Call
Call
myGroup.SynchWrite()
myGroup.AsyncWrite()
Reply
Reply
myGroup_AsyncWriteComplete()
The OPC interface accesses only groups, not individual items.

56. OPC DA: Communication paradigm
shared memory: newer value overwrites the older one
No guarantees
- for same snapshots on different clients
- changes may be missed depending on polling period
OPC DA Client
OPC DA Client
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.
server does not guarantee that different clients see the same snapshot of the plant.
server does not guarantee that all changes to variables are registered, changes may be missed if the polling period is too long.
OPC DA Server

57. 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?
a) Loads symbol table from PLC b) it keeps a record of each client and its groups
Read: sync, async, on change Write: sync, async
Client asks for groups (can define what belongs to them by itself), server depends on proprietary protocol
Can be done event-based, depends on application
Developer must no rely on assumption of consistent snapshot of plant

58. OPC Alarms and Events 4.2.1 OPC 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

59. Alarms and Events: Purpose
e.g.
production start
operator pressed button
OPC AE server registers alarms and events and makes them available to clients
Event: represents change.
Alarm: state machines, may require acknowledgement.
Services offered by OPC AE server
browse 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
e.g.
handle low tank level
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
Alarms and events must not get lost!
Wandida:

60. AE: communication paradigm
message passing: events kept in a queue until all clients have read it (or timed out).
guarantees different clients will see all events in same sequence.
OPC AE Client
OPC AE Client
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 Server
12:
12:

61. AE: Displaying Alarms and Events
Event list can become long (~ 1000 entries), entries are not cleared when the source of the event returns to normal
Alarms list is short (< 50 alarms) appearance changes when alarm is acknowledged, an alarm line is cleared when alarm signal is cleared (but remains in log).
Ack checkbox
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).

62. 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)

63. 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 name
event
event
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)

64. AE: Events - Notification
Tank1LevelHigh_SimpleEvent (source, timestamp, message, severity, category)
AE Client
specified communication
COM/DCOM
OPC AE Server
queue
unspecified communication network, fieldbus or internal bus
event notification
message
Controller
timestamp
Event FB
Tank1
Plant
Level Switch

65. AE: Events - Time Stamp Events can be time-stamped at
- device that produced the data allowing Sequence-Of-Events with a high accuracy (microseconds)
- controller giving accuracy not greater than the period of PLC tasks, about 1 ms.
- OPC Server, when event message arrives not more accurate than DA, about 10 ms
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)

66. AE: Alarms - Condition Definition
Described in a named state machine.
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 > 100 AND FIC101.PV < 150)
Condition
Acknowledgement signal (a positive transition of a boolean variable)
Condition state
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
Enable (positive transition) Disable (positive transition)

67. AE: Alarms - Acknowledgement
Alarm condition active when PLC produces alarm signal for abnormal state.
Operator has to acknowledge this condition (client ack) Alternatively, a button at the PLC may be pushed (field ack)
event notification
Tank1Level_ConditionEvent
AE Client
client ack (acknowledger ID)
COM / DCOM
OPC AE Server
Network, field bus, or internal bus
message
Condition
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)
time- stamp
LevelHigh
controller
AckButton (field ack)
Alarm Signal

68. AE: Alarms - Condition states and acknowledgement
event notification
alarm signal
acknowledgement
Operator
Inactive Acked
Active Unacked
Active Acked
Inactive Acked
Active Unacked
Inactive Unacked
Inactive Unacked
Inactive Acked
condition state
alarm_signal 
condition state transitions
(here: always enabled)
Enabled Inactive Acked
alarm_signal 
EnabledActive Unacked
EnabledActive Acked
Ack 
Ack 
alarm_signal 
Enabled Inactive Unacked
alarm_signal 
Event is generated when alarm signal changes state, or is acknowledged by operator

69. AE: Summary alarms and events
AE Client
AE Client
OPC AE Server
OPC AE Server
event notification (source, timestamp, message)
alarm notification (source, timestamp, message, condition, subcondition, severity, type)
message
Condition
timestamp
Event FB
message
controller
controller
event
alarm
ack

70. OPC A&E: Assessment
What is the difference between an alarm and an event?
Where are alarms and events time stamped?
How does the “message passing” paradigm influence the OPC client application developer?
E: state change, A state machine for abnormal states, can trigger event notifications
Depends, accuracy varies
Ensures the same sequence of events is seen by all clients

71. Usage and specifications Clients and Servers: configuration
OPC Common Overview
4.2.1 OPC 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

72. 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:
Sub-Second Data Collection Rates
Compression
Data collection, archiving and retrieval
Report generation

73. HDA: Historical Data Access
HDA Clients
e.g. Trend Analysis
e.g. Event Logger
OPC HDA Server
hidden
calculations
history
database
raw and ordered data
collector
OPC DA Server
proprietary data acquisition
Field device
Field device

74. HDA: Purpose
Server: access to a historical data base (logs) data from process have been collected and time-stamped
Clients: require ordered access to data e.g., trend analysis, product tracking or data mining
Services
- browse 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
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

75. HDA: Raw log Registers value changes For each item: itemID value
:40
Gpcpt2ofpbonne
4824
Registers value changes
:40
Cpt2bac 50
:40
Gpcpt2bac 70
:40
Gpcptbe2 45
:41
Gpcpt1bac
151
:41
Gpcpt1ofpbonne
4826
For each item:
itemID
value
quality (good, stale, bad)
timestamp (UTC).
:41
Gpcptae2 45
:41
Cpt1bac 49
:41
Gpdefr2 64
:41
Gpvoydef 2
:41
Gpr3tempscycleprd
318
:42
Gpstn1e1 16
:42
Gpalarme1
:42
Gpalarme2
:43
Gpetatmodemarche 2
:43
Gptpscycle
1346
:43
Gpetatmodemarche 1
:43
Gpdefgene1 16
:43
Gpetatmodemarche
:43
Gptpscycle
317
:43
Gpdefr2
By definition, values are registered when they change. (even if data are acquired by periodic polling)
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).
:43
Gpvoydef
:43
Gpdefgene1
:44
Gpetatmodemarche 1
:44
Gpr2tempscycleprd
1992
:44
Gptpscycle
435
:44
Gpalarme3 1
:44
Gpalarme4 1
:44
Gpalarme3
:44
Gpcpt2ofpbonne
4823

76. HDA Application: Trend Display
interpolation
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

77. HDA: Hierarchical logs
long-term log
1 hour_forever
 5.2 MB / Year 1 2 3 27 28 29 30 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
september
octobre
october
1 mn_7 days log
 6 MB
yesterday
d-2
d-3
d-4
d-5
d-6
d-7
today
1s_24 hours log:
50 12 B
 52 MB 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
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

78. Real-Life Examples and Exercise
Integration of redundant GE Mark V turbine controllers at a power plant
The controllers tie into the plant’s main DCS, an ABB Advant control system. Process Portal B and MicroSCADA client applications are used to visualize process data from the turbine controllers.
Exercise: Draw schema of system and describe which OPC specifications are used where 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
and

79. Manufacturing Execution System
EPFL, Spring 2017
Manufacturing Execution System

80. Context A Manufacturing Execution System (MES)
tracks and documents all steps of the transformation of raw materials to finished goods
provides information for decision makers to optimize production output.
Runs in real time to enable the control of multiple elements of the production process (e.g. inputs, personnel, machines and support services).
Group control
Unit control
Field
Sensors
& actors A V
Supervisory
Primary technology
Workflow, order tracking, resources
SCADA =
Supervisory Control
And Data Acquisition T
Production planning, orders, purchase 1 2 3 4
Planning, Statistics, Finances 5
(manufacturing) execution
enterprise
administration
Overview
Hierarchy

81. Example Manufacturing Model: Restaurant
Processes, actors, ressources, consumables, flows
authorities
accounting
controller
recipes
menu
table
clients
clients
cashier
owner
clients
table
clients
client
table
cooks
dish washer
waiters
chef
dish washer
waiter
table
suppliers
fresh
food
prepared
food
waste

82. Components of MES Product & Process Engineering Control
Supply Chain Management SCM
Enterprise Resource Planning ERP
Sales & Service Management
Product & Process Engineering
Control
PLC/ Soft Logic
Drives, Motors Relays
Data Collection
Manual Process Control
DCS/ OCS
MES: Integrated Production
Data, Operations
Management Systems, People
Process Management
CAD/CAM
Product Data Management
Customer Relationship Management
E-Commerce
Automation, Instruments, Equipment
Maintenance Management
Dispatching Prod. Units
Performance Analysis
Procurement
E-Auction
Inbound/ Outbound Logistics
Labour Management
Product Tracking
Scheduling & Planning
Quality Management
Resource Allocation
Document Control
Inventory Management
Strategic Sourcing
source: MESA White Paper

83. ISA S95 standard US standard defining terminology and good practices
Delineate business processes from manufacturing processes
Identify responsibilities and functions in Business to Manufacturing and Manufacturing to Manufacturing integration
Identify exchanged information in Business to Manufacturing integration
Improve integration of manufacturing by defining:
Common terminology
Consistent set of models
Establish common points for the integration of manufacturing systems with other enterprise systems
Part 1: Models and Terminology Part 2: Data Structures and Attributes Part 3: Models of Manufacturing Operations

84. Manufacturing Element Examples
InBox
Production
Cell
OutBox
A palette can carry a
product or a lot
production
order
(recipe)
materials,
parts,
energy
product,
waste
Transporter
Palette
Production
Cell
InBox
OutBox
Store
InBox
Production
Cell
OutBox
production
reports

85. Example Workflow
Preparation
of stator
Winding
Welding
Impregnation
Assembly
of rotor
Final Tests
Workflow: path that the product being manufactured takes through several “stations”.
Recipe: sequence of operations that takes place at one particular station.

86. Workflow: Transportation, productivity and inventory waste …
Order Travel
1. Tubes
2. Unprotected
Cores
3. Protection
4. Taping
5. 2ry Winding
6. Short
Circuit Test
7. Pre-Test
8. Protective Taping
9. Positioning
10. Buffer
11. 1ry Winding
12. Taping
13. Buffer

87. … have been vastly eliminated from the factory floor
Order Travel
Positioning
1ry Winding
Pre-Test
Taping
Protective
Taping
2ry Winding
Buffer
Taped Cores

88. Autonomous robots in manufacturing
source: KIVA, Boston

89. ISA S95: 1. Resource Allocation and Status
Guiding what people, machines, tools, and materials do.
Maintain and display status of resources including machines, tools, labour, materials, etc.
Details
manage resources (machines, tools, labour skills, materials, other equipment, documents, … that must be available for work to start and to be completed, directly associated with control and manufacturing.
do local resource reservation to meet production-scheduling objectives.
ensure that equipment is properly set up for processing, including any allocation needed for set-up.
provide real-time statuses of the resources and a detailed history of resource use.
Similar descriptions for dispatching production, data collection, quality management, process management, product tracking, … See IA_ISA95.pptx for more details.

90. Performance Analysis Goal: answer questions of factory owner.
What is the number of good / bad pieces produced: by shift X, in week 20 ? (with / without induced downtime) What is the relation to the maximum ?
What was the average production speed of a unit compared to the maximum ?
How close to the theoretical production capacity is my plant producing ?
What are the N major reasons why the unit is not producing at full capacity ?
How many stops did the unit suffer from ?
What is the availability of my production unit?
What is the efficiency of operator M ? of shift S ?
What is the progression of the OEE (overall equipment efficiency) on a daily basis ?
How does my OEE compares with others ?
How much time is spent loading / unloading the machine ?

91. Conclusion
MES is a business of its own, that require a good knowledge of the manufacturing
process and organization skills.
Simulation tools are helpful to anticipate the real plant behavior
Fits well with buzzwords like “lean manufacturing”, Industrie 4.0, Digitalization

92. Assessment What are the parts of the ISA S95 standard ?
What is workflow optimization ?
Why do manufacturing models exist ?
What KPIs are manufacturers interested in ?