SAGAR BHARDWAJ

EVOLUTION OF DCS
Different stages of evolution of Control system are :
Manual or Field control Pneumatic Single loop control system. Electronic single loop control system Centralized control system Distributed control system.

DISTRIBUTED CONTROL SYSTEM

DEFINATION :A Control System In Which Different Intelligent
To Control The Process And Integrated On

Devices
Same

(Controllers) Are Functionally & Geographically Distributed

Communication Loop.

ADVANTAGES OF DCS SYSTEM

REDUNDANCY:- The System Consists Of Critical Equipments Which Have Been Supported By Active Stand By System Which Results In Increased Reliability. RELIABILITY :- Ability Of A System To Operate Efficiently Under PreDetermined Operating Conditions. ADAPTABILITY:- Ability Of A System To Operate Under Adverse Operating Conditions Accept Changes And Also Provide Easy Machine Human Interface.

ARCHITECTURES OF DCS
The master less deterministic type on single network (Node Server) The client server type with master less characteristic

on single network
The client server type with server as master and

interlink between clients on two separate networks.

NODE SERVER
A network with deterministic master less type of protocol has a real master less node hierarchy, which means that all the nodes, i.e., all controllers as well as operator stations will have the same priority and shall be always communicating to the network together. In this case, the failure of any one operator station, i.e. any one node will lead to the non operation of that node only and the rest of the system stays healthy , thereby maintaining the healthiness of the network.

Node 1

Node 2

Node 3

Node 4

Bus1 Bus2 All healthy nodes and all are at same hierarchy
Node 5 Node 6

NODE SERVER MASTERLESS TYPE DCS

Node 1

Node 2

Node 3

Node 4

Bus1
failure

Bus2

Node 5

Node 6

All healthy nodes and all are at same hierarchy

Failure of Bus-2 causes Bus-1 to take over without making any difference to nodes

NODE SERVER MASTERLESS TYPE DCS

failure

Bus1 Bus2

FAILURE OF A NODE WILL LEAD TO THE FAILURE OF THE PARTICULAR NODE ONLY. OPERABILITY IS INTACT FROM ALL OTHER NODES

Still Healthy Nodes keep system healthy

NODE SERVER MASTERLESS TYPE DCS

CLIENT SERVER WITH SINGLE NETWORK

A special case of Client-server architecture with servers and all its clients connected directly to the same network. In this architecture both the controllers and the operator stations are clients to the server but communicating directly to the network. All the data required are being stored in a common distributed database at the server which keeps the complete history of the operations and process data.

CLIENT SERVER WITH SINGLE NETWORK

In normal operation all the operations and process data are being communicated from/to the operator stations/ controllers are routed through the server, but failure of the server will lead to a direct communication of the operator stations to the controllers. The complete controllability of the controllers from the operator stations stays intact and the only difficulty faced is the loss of the historical trending capability of the control system.

Server

Client 1

Client 2

Client 3

Bus1 Bus2

Client 4

Client 5

Server healthy means all nodes (clients) communicating to the server

CLIENT-SERVER ARCHITECTURE TYPE DCS BUT MASTER LESS FOR OPERABILITY

server

Client 1

Client 2

Client 3

Bus1
failure

Bus2
Server healthy means all nodes (clients) communicating to the server

Client 4

Client 5

Failure of Bus-2 causes Bus-1 to take over without making any difference to nodes

CLIENT- SERVER ARCHITECTURE TYPE DCS BUT MASTERLESS FOR OPERABILITY

Server

Client 1

Client 2

Client 3

Failure

Bus1 Bus2

Client 4

Client 5

Failure of server renders failure of communication between clients and server

Even after failure of server, the clients keep communicating in between and hence operability is intact

CLIENT SERVER ARCHITECTURE TYPE DCS BUT MASTERLESS FOR OPERABILITY

CLIENT SERVER WITH TWO SEPEATE NETWORKS

A network with client server architecture will have servers as an interconnection point between the controller nodes and the human machine interface nodes (HMIs). Both the controllers and the HMIs are clients to the server. The failure of the server will lead to a highly dangerous situation of complete loss of controllability of the network. No process information from the controller will be communicated to the operator station and instructions from the operator station also cannot be communicated back to the controller.

Client 1

Client 2

Client3

server

Network I /Bus1

Network I /Bus2

Network II /Bus1 Network II /Bus2 Client 4 Client 5

CLIENT SERVER ARCHITECTURE TYPE DCS WITH SERVER AS MASTER

Client 1

Client 2

Client3

server

Network I /Bus1
failure

Network I /Bus2

Failure of Bus-1 causes Bus-2 to take over without making any difference to nodes
Network II /Bus1 Network II /Bus2 Client 4 Client 5

CLIENT SERVER ARCHITECTURE TYPE DCS WITH SERVER AS MASTER

Client 1

Client 2

Client3

Data lost

server failure

Data lost

Failure of the server will lead to failure of complete network and loss of operability

Client 4

Client 5

STEAM GENERATION PLANT OF NFL PANIPAT

Steam generation plant has three Boilers of BHEL make, each having MCR (Max. Capacity Rating) of 150 T/hr respectively, producing superheated steam at a Pressure of 100 Kg/cm2 & at a temperature of 500 0 C All the three boilers are controlled by DCS.

SGP DCS SYSTEM

Distributed control system for steam generation plant is being procured from M/s ABB . The system is known as

WORKSTATION BASED SYMPHONY HARMONY INDUSTRIAL IT SYSTEM.

The control and monitoring philosophy has been split in to two categories i.e. closed loops and Open loops ( mainly indications). The closed loops are being powered , monitored and controlled in the DCS whereas the open loops are being powered and configured in the MUX system and monitored in the DCS system.

SGP DCS SYSTEM NODE DISTRIBUTION

There are total 10 nodes in system which are directly connected to C-NET. Sr.No 1 No. of nodes Three (3) Used for Three Boilers

2
3 4

One(1)
One(1) Five(5)

Common section
EWS OWS

The EWS & OWS are also to Ethernet switch through O- NET cable.

Two printers are on O-NET.

SGP DCS SYSTEM PANEL CONFIGURATION

DCS system panels for each boiler comprises of four panels, one is system panel, and the rest three are termination panel. AI terminates through HMS panel whereas AO terminate directly , DI/DO terminate through relays. System panel have three set of controllers in redundant,one for OLCS, one for CLCS & last for MUX. System panel communicates using NIS & NPM. Similarly OWS communicate on C-NET using NIS & ICT.

DATA FLOW

O-NET
NODE-9 NODE-8 NODE-7 NODE-6

LINE-1 LINE-2

NODE-10

NODE-5

ITC

ITC

ITC

ITC

ITC

ITC

NIS

NIS

NIS

NIS

NIS

NIS

REDUNDENT C - NET
NIS NPM
NODE-1 BOILER-1 BOILER-2 NODE -2 BOILER-3 NODE -3 COMMON

NIS NPM

NIS NPM

NIS NPM
NODE-4

C-NET
C-NET is a unidirectional, high speed serial data network that operates at a 10 megahertz communication rate. It supports a central network with up to 250 system NODE connections.

SYSTEM CARDS
THERE ARE DIFFERENT TYPES OF CARDS IN SYSTEM PANEL FOR VARIETIES OF INPUTS/OUTPUTS1)TEMPERATURE CARD FOR THERMOCOUPLE/RTD,( IMAS123)
2)ANALOG INPUT CARD FOR 4-20 ma, (IMFEC12) 3)ANALOG OUTPUT CARD FOR 4-20 ma(IMASO11) 4)DI CARD FOR CONTACT INPUT (IMDSI 13) 5) DO CARD FOR CONTACT OUTPUT (IMDS0 14) 6) NETWORK INTERFASE SLAVE (NIS-21) 7) NETWORK PROCESS MODULE (NPM-12) 8) BRC 300 ( CONTROLLER) 9) MFP 12 ( CONTROLLER) 10) ITC- (FOR COMMUNICATING WITH THE COMPUTER)

DCS SYSTEM
4-20 mA out of smart Txs., Superimposed with hart signal, are terminated to HMS. Here, HART Signal and 4-20 mA signals are separated using patch cards. 4-20 mA goes to terminate to AI cards through DCS Marshelling cabinet. Termination of this 4-20 mA depends on whether this is field powered or system powered.

MULTIPLEXER SYSTEM
Supplied by MTL,England. Multiplexers are used for cable saving. Here non critical signals have been taken through MUX. There are four MUX stations for each boiler installed in field itself. All the twelve MUX stations having redundant supply . MUX -1 & 2 for each boiler caters mainly AI whereas MUX-3 & 4 caters mainly DI/DOS.

MULTIPLEXER SYSTEM
The MUX uses redundant EBIM (Ethernet bus interface module) processing data at very high speed of 100 Mbps, whereas the controller (MFP12) used in DCS system for third party communication is Modbus compatible, hence MOXA switches are used to convert Ethernet to Modbus.

MULTIPLEXER SYSTEM
Hence redundant Ethernet cable have been laid from each of the MUX to redundant MOXA switches from MOXA, Ethernet Cable been laid for HMS/MUX PC where WORKBENCH Software is loaded for status monitoring of all MUX modules. Also RS-485 Cable laid from MUX-1 & 2 having AIs only to HMS PC, through RS-485 to RS-232 Converter, for catering smart transmitters with the help of CORNERSTONE Software loaded on the same HMS/MUX PC.

SER-SEQUENCE OF EVENT
RECORDER
There is one SER of 512 points comprising of two SER Stations - EL1 & EL2. SER Stations EL1 DI Cards 9 No of events handling capacity 288

EL2

224

DATA FLOW BETWEEN BRC AND C NET

C-NET C-NET

NIS

NPM

TO FIELD

TU

SLAVE

BRC

FROM FIELD

TU

SLAVE

BRC

CAPTIVE POWER PLANT IN NFL PANIPAT

In NFL Panipat We have Captive power plant which has MCR 230 T/hr at Pressure of 100Kg/cm2 and at a temperature of 500oC In NFL Panipat we have Two Turbo Generators (TGs) each having a capacity of 15MW Boiler of CPP and Turbine of TGs is controlled by DCS.

CPP DCS SYSTEM

1. Distributed control system for CPP was originally procured from

M/s ABB The system was BAILEY NETWORK 90

2. Later on it was up graded with Symphony Harmony Industrial IT System for turbine side 3. The boiler side is up graded with INFY 90 System. 4. In this way CPP have to separate DCS System. 5. Both systems are connected to a pair of redundant server. 6. CPP DCS System is Client server based Architecture.

NODE DISTRIBUTION OF CPP DCS SYSTEM

1. It Has 16 (Sixteen) Nodes connected to C NET

Sr.No. 1 2 3 4 5 6 7

No. of Nodes Three(3) Four(4) Three(3) Two(2) One(1) Two(2) One(1)

Used for Automatic Boiler Controller(ABC) Burner management system(BMS) Sequence (SEQ) Turbine of TG 1 & 2. Common Section RTDS 1 & 2 EWS

OPERATING STATIONS DISTRIBUTION OF CPP DCS SYSTEM

Four Operating Station Are Connected Through

Redundant O Net To RTDS 1 & 2.

A. Two Stations Are Loaded with History. B. One Station Is Loaded with Configuration Software. The Operating Station System Is Windows 2000

Professional Based.

ARCHITECTURE OF CPP DCS SYSTEM

ABC SEQ BMS TG-1 TG-2 COM

C - NET

RTDS1

RTDS2

EWS

O-NET

HIS P

HIS R

CLIENT

COMPOSER

COMPARISON OF DCS OF SGP & CPP OF NFL PANIPAT

SGP
Architecture is Node Server based. Operating system is Window XP Professional based. All critical loops are redundant DCS is PGP based

CPP
Architecture is Client server based. Operating system is Windows 2000 professional based. All critical loops are not redundant . DCS is PPB based.

DCS COMPARISON OF VARIOUS MODELS OF VARIOUS COMPANIES

COMPANY S.N O. MODEL Advant with 460 controller ABB Syphony Harmony IT Infi-90 CS 3000 YOKOGAWA CS 1000

1 2

HMI Architecture Operating System

Node Server option Window Based Window NT

Node Server option Window Based Window NT or XP Using C-Net 10 Mbps

Client Server Window Based Window NT Using C-Net 10 Mbps

Node Server Window Based Window NT,XP,2000 Using V-Net 10 Mbps

Client Server Window Based Window NT,2000 Using VL-Net 10 Mbps

3 4

Data Bus Redundancy A) Power Supply B) System Bus C) Controller D) I/O Module E) Server F) Memory(system )

Using C-Net 10 Mbps

Yes Yes Yes optional with extra Yes

Yes Yes Yes Present in critical loop Yes

Yes Yes Yes optional with extra Yes

Yes Yes Yes Present in critical loop Yes

Yes Yes optional with extra optional with extra Yes

Yes

Yes

optional with extra

Yes

optional with extra

DCS COMPARISON OF VARIOUS MODELS OF VARIOUS COMPANIES

COMPANY S.N O. 5 MODEL Processor A) System B) Controller C) Controller Memory User Defined User Defined User Defined User Defined VR5432(133MHz) 32Mb User Defined Proprietary Advant with 460 controller ABB Syphony Harmony IT Infi-90 CS 3000 YOKOGAWA CS 1000

68040/25 MHz, Motorola

Up to 8Mb

32- bit, Motorola RISC

16Mb

32- bit, Motorola RISC

16Mb

6 7
8 9 10 11

Online Replacement Online Configuration System Memory I/O capacity Web Connection Scan time<200ms

Yes Yes
User Defined >10000 tags posible Possible with isolation Possible

Yes Yes
User Defined Sutable for large plant Possible with isolation Possible

Yes Yes
User Defined only for medium plant Possible with isolation Possible

Yes Yes
User Defined Up to 100000 I/O Possible with isolation Possible

Yes Yes
User Defined only for medium plant Possible with isolation Possible

DCS COMPARISON OF VARIOUS MODELS OF VARIOUS COMPANIES

COMPANY S.N O. MODEL FOXBORO IA Series HONEYWELL Experience PKS TPS with gus Teleperm XP SIEMENS Telemerm ME

1 2

HMI Architecture Operating System

Node Server option Window Based Window NT,XP,2000

Client Server Window Based Window NT,XP,2000 Using LAN Up to 100 Mbps

Node Server option Window Based Window NT Using LCN 5 Mbps

Node Server Window Based

Client Server DOS Based STRUK UsingCS-272 <10Mbps

3 4

Data Bus Redundancy A) Power Supply B) System Bus C) Controller D) I/O Module E) Server F) Memory(system)

Using Device-Net 10 Mbps

Yes Yes Yes Present in critical loop Yes Yes

Yes Yes Yes optional with extra Yes optional with extra

Yes Yes optional with extra optional with extra Yes Yes

Yes Yes optional with extra optional with extra Yes Yes

Yes Yes optional with extra optional with extra optional with extra optional with extra

DCS COMPARISON OF VARIOUS MODELS OF VARIOUS COMPANIES

COMPANY S.NO . 5 MODEL Processor A) System B) Controller User Defined N/A User Defined C 200 4000Kb Yes Yes Yes Yes User Defined N/A User Defined Virtual Processor INTEL PII INTEL 80188(20MHz) 156Kb Yes FOXBORO IA Series HONEYWELL Experience PKS TPS with gus Teleperm XP SIEMENS Telemerm ME

C) Controller Memory
6 Online Replacement Online Configuration

7 8
9 10 11

Yes User Defined

Up to 100000 I/O Possible with isolation Possible

Yes User Defined

Up to 100000 I/O Possible with isolation Possible

Yes User Defined

10000 tags Possible with isolation Possible

Yes User Defined

Yes User Defined

only for medium plant

System Memory
I/O capacity Web Connection Scan time<200ms

Possible with isolation Possible

N/A Possible