Old Personal Computer Hardware SERVERS

• Operating Systems
  • Windows
  • Windows Server 2008 R2 (July 2009)
  • Windows Server 2008 (February 2008)
  • Windows Server 2003 (April 2003)

  

• Network Cards (NIC – Network Interface Card)
  • The first thing to check is the number of ports
  • Typically it’s either single or dual port
    • Meaning how many Ethernet cables can go in
  • The next thing to check is it’s connection
    • If it say PCIe only that means it is PCIe x1 and compatible with all PCIe slots (X1,X4,X8,X16)
  • Another thing to check is it’s bandwidth, they are:
    • Ethernet or 10BASE-X
      • Only supports up to 100Mbps or 1.25MBps
      • Old Technology
    • Fast Ethernet or 100BASE-X
      • Support up to 100Mbps or 12.5MBps
      • Old Technology
    • Gigabit Ethernet or 1 GbE
      • Supports up to 1000bps or 1Gbps = 125MBps
      • Backward compatible
      • Can use same CAT5 Cables and connect to 10/100 hardwares
        • However, it causes the new equipment to be slower
      • 1GBe Ethernet has 5 Varieties in physical layer standards which are

Name	Medium	Specified distance
1000BASE‑CX	Twinaxial cabling	25 meters
1000BASE‑SX	Multi-mode fiber	220 to 550 meters dependent on fiber diameter and bandwidth[2]
1000BASE‑LX	Multi-mode fiber	550 meters[3]
1000BASE‑LX	Single-mode fiber	5 km[3]
1000BASE‑LX10	Single-mode fiber using 1,310 nm wavelength	10 km
1000BASE‑ZX	Single-mode fiber at 1,550 nm wavelength	~ 70 km
1000BASE‑BX10	Single-mode fiber, over single-strand fiber: 1,490 nm downstream 1,310 nm upstream	10 km
1000BASE‑T	Twisted-pair cabling (Cat‑5, Cat‑5e, Cat‑6, or Cat‑7)	100 meters
1000BASE‑TX	Twisted-pair cabling (Cat‑6, Cat‑7)	100 meters

  

• TOE (TCP Offload Engine)
  • New Network cards especially the 1GbE types have TOE (TCP offload engine). What this does is that it has a processor which processes various TCP protocols such as the 3-way handshake which is supposedly to be done by the processor.
  • The purpose is to free up processor speed

  

• Hard Disk
  • Interface
    • Hard disk commonly interfaces with the PC using a serial bus interface.
    • The most common interfaces:
      • SATA
        • Cheaper because it does not use a built in processor for controlling head position unlike SAS
      • SAS (Also known as SCSI)
        • Faster and more reliable
    • Some Hard disks have SED feature (Self Encrypting Device). This is for security purpose
  • Hot plug
    • A property of a hard disk which can plugged out when the server is running. This avoids server downtime when a hard disk change is required (due to faulty hard drive and etc). Hot plug hardwares cost more
  • RAID
    • Raid stands for (Redundant Array of Inexpensive Disks). The advantages of RAID
      • High Data Reliability and Availability
      • Ease of combining hard disk space
      • Improve Speed
    • Raid Controllers
      • When RAID is selected, a RAID controller needs to be installed. For dell, this is called the PERC (PowerEdge Raid Controller).
      • Types of RAID Controllers:
        • Software RAID Controller
        • Hardware RAID Controller
    • RAID Numbers:
      • 0 (Stripping)
        • 2 Hard disks mirrors each other not for redundancy.
          • It Doubles data transfer rate, theoretically, but in real life the gains are around 10%
          • Improves Performance
        • Minimum 2 Drives
        • 
      • 1 (Mirroring)
        • 2 Hard disks mirrors each other
        • Advantage
          • Highest Reliability
        • Disadvantage
          • Size is limited by the drive with smallest size
        • Minimum 2 Drives
        • 
      • 2 (Bit Level Stripping with Hamming Code parity drives)
        • Improve data integrity and reliability
        • Obsolete
        • Minimum 3 Drives
      • 3 (Byte Level Stripping with dedicated parity drive)
        • Hardly Used
        • Minimum 3 Drives
        • Same advantage
      • 4 (Block Level Stripping)
        • Hardly Used
        • Minimum 3 Drives
        • 
      • 5 (Block level stripping with distributed party)
        • Commonly Used
        • Advantages
          • Provides Redundancy up to one hard disk failure
          • Increase Disk Size to (DriveSize x (Number of Drives – 1))
        • Disadvantage
          • If more than 1 hard disk fail, ALL data are LOST
        • Minimum 3 Drives
        • 
      • 6 (Block level stripping with double distributed parity)
        • Needs at least 4 drives
        • Tolerates up to two failures
        • Minimum 4 Drives
      • 10 (Mirroring without parity and block level stripping)
        • Is a combination of RAID 1 and RAID 0
          • Works in Spans
            • Each span will mirror one another (Like Raid 1)
          • The upper span works like RAID 0
        • Advantage
          • Speed and Reliability Combined
        • Disadvantage
          • Less disk space compared to RAID 5
        • 

• Motherboard / Chipset

• DTE/DCE
  • DTE – Data Terminal Equipment, Ends a communication line (The processing node)
  • DCE – Data Communications Equipment, in the middle of communication line (The bridge)

• PCI Slots (Peripheral Component Interconnect)
  • Conventional PCI
    • 1991
    • 33/66 MHz
    • 6 types of slots
      • 3.3V 32 Bit
      • 5V 32bit
      • 3.3V 64 Bit
      • 5V 64 bit
      • Universal 32 bit
      • Universal 64 bit
    • 32 bit PCI slots can be used in 64 bit slots

• PCI-X
  • 1999
  • 66/133 Mhz
  • 3.3V 64 Bit
  • Faster than Conventional PCI
  • Conventional PCI cards can be used on PCI-X slots
  • Supports Multi Drop

• PCI Express / PCIe
  • 2002
  • Formerly known as 3GIO
  • Have 4 types of slots
    • X1, X4, X8, X16
  • The smaller slots can fit in larger slots. That is why some hardware just specify itself as PCIe which means it is PCIe x1 and can fit in all slots
• A typical motherboard will have these slots

• Server Rack
  • The most common server rack is the 19” rack. It is specified in EIA , CEA and IEC standards
  • Typical rack servers are around 17 to 18”. The racks are mounted by a slide

• Ethernet
  • Is a physical layer LAN technology which is defined in the IEEE standards. LAN is a network limited by geography
  • The standards
    • IEEE 802.3 (Also called Wired Ethernet or Simply referred to as Ethernet)
    • IEEE 802.11 (Commonly referred to as Wifi)
    • IEEE 802.16 (Commonly referred to as Wimax

• RAM (Random Access Memory)
• What is RAM?
• RAM is a type of Computer Storage. The word random-access means that data can be read anytime without going through a sequential process.
• However, today this has become a name only, modern DRAMs use burst mode reading which is sequential in nature but fast. SRAM however, still works in a RANDOM access way
• RAM are Volatile
• When power is lost, data is lost in the RAM.
• Static (SRAM) vs Dynamic (DRAM)
• SRAM is faster and more expensive. It is often used as cache memory for the CPU
• Don’t need to know much about this other than when buying a CPU.
• DRAM is slower and typically used as computer memory.
• DRAM is the most common RAM Type.
• DRAM Formats (these are just how the memory is designed)
• SIMM (Single Inline Memory Module)
• Old Technology
• 32Bit
• RIMM (Rambus inline Memory Module)
• Old Technology between SIMM and DIMM
• DIMM (Dual Inline Memory Module)
• MOST COMMON FORMAT
• Typically around 4-5 Inches length
• Replacement Technology for SIMM
• 64Bit
• Twice more Pins than a SIMM, Faster than SIMM
• The number of Pins in DIMM is determined by the specification type (e.g. DDR2, DDR3, SDRAM)
• SDRAMs (The old type) -168 PINS
• DDR1 (OR DDR SDRAM) -184 PINS
• DDR3 and DDR2 -240 PINS
• SO-DIMM (Small outline dual in-line memory module)
• Used is system with limited space such as Laptops
• Typically 2.5-3 Inches Length
• Number of Pins
• DDR2 – 200 Pins
• DDR3 – 204 Pins
• DDR4 – 256 Pins
• DRAM Asynchronous vs Synchronous
• Asynchronous DRAM
• Not suitable for high speed systems
• Synchronous DRAM
• Since it’s synchronous, it waits for Clock Signal. This is why Speed is rated in MHz rather than in Nano seconds
• Common types of DIMM
• SDRAM (Synchronous Dynamic Random Access Memory)
• First Generation, Slower than DDR
• Write 1 Words per clock cycle
• DDR SDRAM [MOST COMMON TYPE WHEN OPENING ON A LAPOP OR DESKTOP]
• DDR Stands for Double Data Rate
• Unlike SDRAM, it write 2 words per clock cycle
• Evolution of DDR SDRAM
• DDR1 (Also just called DDR SDRAM)
• DDR2
• DDR3
• DDR(X)’s are not backward or forward compatible with each other, meaning if a mother board supports DDR1 it will not Support DDR2 and Vice Versa
• DDR Standard is set by JEDEC
• On a Memory, a typical annotation would be
• DDRx-yyyy – x is the DDR version, yyyy is the data rate of the memory measured in MT/s (Mega Transfer Per Second).
• This is essentially the Clock Speed which is usually advertised.
• PCx-yyyy – x is the DDR Version, yyyy is the peak transfer rate measured in MB/s. MB/s = MT/s x 8
• For example, PC2-6400 means it is a DDR2 Ram with Speed of 6400/8 = 800 MHz
• Brands
• Good Ones
• Crucial
• Corsair
• Kingston
• G Skill
• Bad Ones
• RAM Clock Speed
• Should generally choose the highest  RAM Speed supported by the mother board
• One should not use RAM of different speeds as this causes instability
• Number of PINS
• This is important when purchasing a RAM as the number of PINS need to match the motherboard
• Typically the following rules apply:
• SDRAMs (The old type) have 168 PINS
• DDR1 (OR DDR SDRAM) has 184 PINS
• DDR3 and DDR2 has 240 PINS
• Single or Pair?
• Pair is better as it allows Dual Channel
• 2x2 GB RAM is better than 1x4GB RAM

• Power Supply Unit (PSU)
• A computer PSU’s job is to convert mains AC to low-voltage regulated DC power
• Specification Standards
• ATX
• Most Common
• BTX
• Number of Pins
• The number of Pins required would depend on the motherboard
• 24 Pins
• Connectors
• Typical Power supplies have several connectors, and this is usually standard
• P1 - PC Main Power Connector
• This goes to the Motherboard
• Has either 20 or 24 Pins
• Power supply with 24 Pin connector can be used on motherboard with 20 pin connector
• Some 20 pin connectors have an additional 4 pin connector (referred to as 20+4), which can be put side by side to create a 24 pin connector
• 4 Pin Peripheral Power Connector
• Goes to various disk drives
• Black wire is ground
• Yellow is 12v, Red is 5v
• Can be any label
• Serial ATA Power Connectors
• Used for Hard drives and Optical Drives
• 15 pins inside
• Can be of any label (P2 to p15)
• Floppy Power Connector
• For floppy drives
• Is a small connector with a protrusion on the pin between red and black
• 4 pins – yellow(12v), black (ground), black (ground), red (5v)
• Can be of any label (p2 – p15)
• ATX 4-Pin Power supply connector
Used to provide 12VDC to the processor voltage regulator, which is on the mother board.
• Can be of Any label (p2 – p15), typically p2

Alarm Management

• Definitions of an alarm
• Based on ISA 18.2 - An audible and/or visible means of indicating to the operator an equipment malfunction, process deviation, or abnormal condition requiring a response
• The keywords
• Audible and/or Visible Indication
• For the operator
• Indicating either
• Equipment Malfunction
• Process deviation
• Abnormal condition
• Requiring Response
• The Core Principles in an Alarm System
• Relevant (Not Spurious)
• Prioritized
• Informative (With documentation and diagnostic)
• Unique (Not Redundant)
• Timely
• Advisory (Action Required)
• Purpose of Alarm is to prevent
• Personal injury
• Environmental
• Economic Loss
• Equipment Damage
• Product Quality
• Downtime
• Benefits of good alarms system
• Improve Operational Effectiveness
• Avoid costly plant trips
• Without effective alarming, demand is on ESD
• Minimize upset time
• A good alarm system quickly alerts operator to bring the plant to stable state
• Improve HSE
• HSE related alarms such as fire alarms, environment and occupational safety alarms
• Reduce operator’s load
• More time to concentrate on optimization
• Legal and Insurance Compliance
• Alarm Screening / Identification
• A pre-rationalization process using a combination of maintenance, sophisticated analysis tool and experience to quickly minimize the number of alarms
• Performed by Control System Engineer, E&I Engineers, Operators and Technicians
• Alarm Rationalization
• The process of reviewing each alarm identified from the identification stage
• Purpose of rationalization is to identify
• Purpose and validity
• Consequence of inaction
• Alarm Class
• Priority
• Alarm Set point
• Operator Action
• Alarm Rationalization Alarm Checklist. There should be an alarm only if :-
• There is a potential cause of the alarm
• There is need and possibility for operator intervention
• There is a consequence of inaction
• Consequence of Inaction
• Is typically a set of tables which are divided into
• Economic
• Personal Safety
• Environment
• Reputation (Normally not included)
• Economic Impact can be measured in
• Time
• Cost
• % (of operating cost / feed)
• Lost in production unit (such as Mtpa for mining plant, BPSD for oil refinery and etc)

The details of these definitions are outlined below:

· Employee / Contractor / Personnel Safety	· To estimate the personnel consequence of an event, consider the following extensions of the keywords given in the margin of the risk matrix
· Low	· Moderate	· High
· Reversible health effects of concern. · Reversible injuries requiring treatment, but does not lead to restricted duties. · Medical treatment.	· Severe reversible health effects of concern. · Reversible injury or moderate irreversible damage or impairment to one or more persons. · Lost time illness or injury.	· Life threatening or irreversible health effects or disabling illness. · Single fatality and/or severe irreversible damage or severe impairment to one or more persons.

Table 1 – Personnel and Safety

 

· Environment	· To estimate the personnel consequence of an event, consider the following extensions of the keywords given in the margin of the risk matrix
· Low	· Moderate	· High
· Near-source confined and short-term reversible impact. · (Typically a week)	· Near-source confined and medium-term recovery impact. · (Typically a month)	· Impact is unconfined and requiring long-term recovery, leaving residual damage. · (Typically years).

Table 2 – Environmental Impact

 

· Reputation / Community Trust	· To estimate the personnel consequence of an event, consider the following extensions of the keywords given in the margin of the risk matrix
· Low	· Moderate	· High
· Impact on reputation of a Business Unit. Significant public exposure in local media. · Tangible expressions of trust / mistrust amongst a few community members with some influence on public opinion and decision-makers.	· Impact on reputation of Product Group. Comment from national NGO which impacts credibility with neighbours/ regional government. Public exposure in the national media. · Tangible expressions of trust / mistrust amongst some community members with moderate influence on public opinion and decision-makers.	· Impact on reputation of Rio Tinto Group. Comment from international NGO. Public exposure in international media. · Tangible expressions of trust / mistrust amongst most community members with significant influence decision-makers. Widespread loss / gain of trust across the community setting the agenda for decision-makers and key stakeholders.

Table 3 – Community & Reputation

 

· Business / Production Loss / Equipment Damage	· To estimate the personnel consequence of an event, consider the following extensions of the keywords given in the margin of the risk matrix
· Low	· Moderate	· High
· < 2.5% of Operating cost · < 0.15 Mtpa · < 4hrs Downtime	· 2.5 – 7.5% of Operating cost · 0.15 – 0.5 Mtpa · 4hrs – 8hrs Downtime	· >7.5% of Operating cost o 0.5 Mtpa · > 8hrs Downtime

Table 4 - Business Impact

 

 

• Alarm Priority
• There are typically 3 Alarm Priorities which determines the type of audible sound and visual on a HMI
• The 3 Priorities
• Low
• Medium / Normal
• High / Critical
• Alarm Priority Matrix
• Is typically a 3 by 3 matrix which has Consequence and Response Urgency dimension
• It is used to determine the priority of an alarm during alarm rationalization

· Urgency of Controller Response	· Low	· Moderate	· High
· > 30 minutes (longest time)	· Priority “Low”	· Priority “Low”	· Priority “High”
· 5 to 30 minutes (typical time)	· Priority “Low”	· Priority “High”	· Priority “Critical”
· < 5 minute (fastest time)	· Priority “High”	· Priority “Critical”	· Priority “Critical”

Alarm Priority Matrix

The example above is for a mining plant which is usually slow response. For process plant, the controller urgency would be >10 minutes, 2-10 minutes, <2 minutes

 

 

• Alarm Priority Distribution
• It is a best practice that priority level is distributed in the following manner
• High = 5-15%
• Medium = 15 – 30%
• Low = 55 – 80%

Alarm Management Reports

• The fundamental alarm management report is the Alarm count over time
• Cluster Analysis
• Used to analyze chattering alarms.
• In Matrikon Alarm Manager Maximum number of events to analyze is the number of events to analyze to check on a cluster. It is good to put the numbr as high as possible such as 10,000,000
• A cluster is a group of alarms that repeats itself within the specified cluster time (subsequent windows). This is also defined as time windows. A time window is typically set as 60 seconds
• From the report,
• The number of cluster tells how many clusters are there.
• The average cluster member tells the average number of time the alarm chatters in the cluster
• From the report, the most important analysis that can be done is the total alarm occurrences vs chatter occurrences; the ratio of this is the cluster member%.
• Total alarm occurrence is the total number of alarms.
• Chatter occurrence is the total number of alarms which is inside a chattering window.
• The cluster member % is the chatter occurrence / alarm occurrence
• Symptomatic Analysis
• Is used to find an alarm that will always tend to appear after a parent alarm occurs. This alarm (often referred to as the child alarm) is said to be predictable.
• The predictability measure the % of occurrence the child alarm comes up after the parent alarm. It is recommended that a predictability > 50% is considered predictable and the child alarm can be automatically inhibited should the parent alarm come out
• The significance measures the % precedence of the parent alarm when a child alarm comes up.
• Total Alarm Count
• Sum of Audible Alarms (exclude alarms that default filters apply to) between the start and end time.
• Total Event Count
• Sum of all Events (Alarms, Return to Normals, Acknowledgements, Operator Actions, System Messages, etc) between the start and end time.  Exclude events that default filters apply to.
• Total Intervention Count
• Sum of all Interventions (message type is Operator Action) between the start and end time. Exclude events that default filters apply to.
• Alarm Rate
• The Alarm Rate represents the average number of alarms per hour for the selected areas (exclude alarms that default filters apply to).  Divide by the sum of the operators assigned to those areas.
• Peak Alarm Rate
• The Peak Alarm rate will divide data into 10-minute slices.  Take the maximum number of alarms in a 10-minute slice for the selected areas (exclude alarms that default filters apply to).  Divide by the sum of the number of operators assigned to those areas.  Multiply the result by 6 for an hourly peak alarm rate.
• Percent Upset (Percent Hours in Burst Mode)
• Slice each 10 minures. If in 10 minutes there are more than 5 alarms, flag that as burst
• Percent Upset refers to the percentage of 10minute chunks where more than 5 alarms "annunciated" per operator. 
• Intervention Rate
• Intervention Rate = Sum of Interventions (message type is Operator Action) for selected areas during the interval / Total Number of operators assigned to those areas / Number of hours in the interval.
• Intervention to Alarm Ratio
• Intervention to Alarm Ratio = Intervention Rate / Alarm Rate
• Priority Distribution
• Priority Distribution represents the condition field as a percentage.  Results should match the "Alarms By Condition Range" in Excel.
• A good priority distribution should be as follows :-
• Critical Alarm = 0
• High Alarms 0-5%
• Medium Alarms 5% - 15%
• Low Alarms >80%
• Top 20 Alarm Percent
• Sum of 20 "Most Frequent Alarms" divided by the total number of alarms in the time span selected.  Filters used should apply to both the numerator and the denominator.
• Top 20 Interventions Percent
• Sum of 20 "Most Frequent Interventions" divided by the total number of interventions in the time span selected.  Filters used should apply to both the numerator and the denominator.
• Average Alarm Rate (10 mins)
• This calculation is performed exactly as Alarm Rate.  Divide the result by 6 for a 10 minute slice.
• Maximum Alarm Rate (10 mins)
• This calculation is performed exactly as Peak Alarm Rate, however do not multiply the final result by 6 as you would in Peak Alarm Rate.
• Average Alarm Rate (Daily)
• Average Alarm Rate (Daily) = Sum(Total # of Audible Alarms) / Total Number of Assigned Operators / # of Days (for less than 1 day use a fraction to represent the number of hours)
• Maximum Alarm Rate (Daily)
• If the Interval is 1 day or less, the Maximum Alarm Rate will be the same as the Average Alarm Rate (Daily).
• If the Interval is greater than 1 day, then:
• Maximum Alarm Rate (Daily) = Sum(Total # of Audible Alarms) / Total Number of Assigned Operators.  Repeat this process for each day in the selected interval.
• Intervention Rate (Daily)
• Daily Intervention Rate = Sum of Interventions (message type is Operator Action) for selected areas during selected interval / Total Number of Operators for those areas / Number of days in the interval.
• Intervention Rate (10 mins)
• 10 min Intervention Rate = Daily Intervention Rate = Sum of Interventions (message type is Operator Action) for selected areas during selected interval / Total Number of Operators for those areas / Number of 10-minute slices in the interval (typically 6).
• Percent Time < 1 Interventions
• Sum of 10-minute intervals where "Intervention Rate (10 mins)" is less than or equal to 1 divided by the total number of 10-minute intervals x 100%.
• This calculation is already normalized per operator.
• Percent Time > 10 Interventions
• Sum of 10-minute intervals where "Intervention Rate (10 mins)" is greater than or equal to 10 divided by the total number of 10-minute intervals x 100%.
• This calculation is already normalized per operator.
• Percent Time 1-10 Interventions
• 100% - (Percent Time < 1 Interventions) - (Percent Time > 10 Interventions)
• Percent Time < 1 Alarms
• Sum of 10-minute intervals where "Average Alarm Rate (10 mins)" is less than or equal to 1 divided by the total number of 10-minute intervals x 100%.
• This calculation is already normalized per operator.
• Percent Time > 10 Alarms
• Sum of 10-minute intervals where "Average Alarm Rate (10 mins)" is greater than or equal to 10 divided by the total number of 10-minute intervals x 100%.
• This calculation is already normalized per operator.
• Percent Time 1-10 Alarms
• 100% - (Percent Time < 1 Alarms) - (Percent Time > 10 Alarms)
• Performance Category
• 

DCS Process Computer System

• DCS (Distributed control system) is a microprocessor based system which gather input from various instruments and HMI, processes the inputs and sends the required output to other instrument
• The key difference between DCS and SCADA
• SCADA is generally used for monitoring only with limited controls.
• The plant cannot run without DCS but can run without SCADA
• SCADA is mainly for remote monitoring
• DCS Network
• DCS Networks exists due to the incapability of a single processor to compute all requirements of a plant
• Yokogawa DCS network is called V-net, it is similar to the ethernet style and is in accordance to IEEE standards
• Has redundancy due to it’s criticality
• 10 Mbps speed
• Must update data every 1 second
• Allows fibre optic communication for long distance communication such as communication with FAR(field auxiliary room) room
• A Far room is required because
• A DCS network is used for communication between
• ICS
• FCS
• CGW (Computer Gateway) or ACG
• A DCS system will also have another network which is of less important. It will be also of Ethernet type. Yokogawa calls it the E-Net Network. The E-Net network communicates between
• EWS (Engineering/Maintenance Work Station). For updating the DCS
• HMI (Human Machine Interfac, (Yokogawa calls it ICS). This is for graphic updates
• Historians such as Alarm PC
• The reason for this E-Net network is to minimize traffic congestion
• The diagram below shows a simple Yokogawa Centum CS network
• 
• FCS Computer
• The FCS is a microprocessor based station which performs all the control algorithms
• The DCS contains many FCS’ connected via a network (such as Yokogawa VNet). This is:
• To distribute the processing load
• To ensure if one FCS goes down, the other units may can still run.
• For centum CS, the AI+AO’s are limited to 640 and DI + DO’s are limited to 4096
• Sub System Communication
• Sub-system communicates with DCS through
• ACG (Communiccation Gateway)
• Communication (Modbus)
• Critical Sub-systems such as IPS are connected directly to the FCS computer. The connection is by a communication card which utilizes RS485 cable. The communication protocol used is modbus. Such systems are:
• IPS (PLC)
• ADAS
• MMS
• CCC
• HIPS
• MOV
• F&G System
• Since FCS’ are scattered it is typical for a single subsystem to send multiple RS485 cables to different FCS’
• Other subsystem
• DCS IO’s
• IO’s are connected through
• Direct connection to FCS
• Remote connection to FCS from and RIO bus which may utilizes fibre optics communication if more than 750m. Fibre optics are used due to it’s long distance reliability and able to send more data
• The picture below shows a typical IO connection to an FCS computer. As can be seen
• The screen cables are looped all the way and grounded at the main distribution frame cabinet (MDF) at the IE point
• SPD’s are installed in front of IS barriers.
• IS Barriers are only required for IS type instruments
• SPD’s are required except for short cable i.e. if the instrument is inside the control room itself
• The wires that goes directly to DCS IO card named AAM11 via an ELCO cable
• Since this is a Smart HART transmitter there is a HART MUX connection from the IS barrier which goes directly to AMS system via RS485 connection cable
• 
• IO modules for DI and DO will come before a relay board. The relays are usually manufactured by Omron
• When an instrument sends a contact, it will hit close the circuit and causes the relay to toggle
• DI and DO’s will have redundant IO cards if the are critical
• Dry contact and wet contact are common terminologies to describe a particular instrument. Instrument here can be a valve, motor, MOV or relay.
• Dry contact means that when the contact closes, no power is supplied to the receiving system. The contact is just a switch
• One example is a relay. A relay is a dry contact if the relay itself when giving contact, does not gives any power to the receiving instrument, such as an MOV. In this case power comes from the MOV. The relay just completes the loop
• Looking from an input expect, a level switch is a dry contact if when the contact closes, the level switch does not send any voltage to the DI relay/DCS. In this case voltage is supplied by the relay panel
• A wet contact is vice versa, power is supplied when the contact is closed
• As a general rule when speaking about dry or wet contact, it is to be referred to in the context of the instrument/Relay. Every connection must have a supplying voltage. If the instrument/Relay supplies the voltage, it is therefore a wet contact else it’s a dry contact. Commonly people will say, “This relay is a wet contact” means that the relay will supply power when it’s energized
• For outputs, relays are commonly Dry since we let the valve/pump (for example) source it’s own voltage
• For inputs, relays are commonly wet, since most DI’s are mechanical e.g. push button, indication and level switch. It is there for more feasible that the instrument let the DCS decide which voltage it prefers to use.
• You may ask, why does a DI sensor input such as a level switch be connected to a relay board? Can’t the DI close a contact which is connected to the DI card of the DCS? Well, the reason is due to standard engineering practice where the DCS system is to be isolated from the input. This is to protect the DCS system from the instrument voltages or short circuits
• The diagram below shows a typical redundant DI loop for a particular instrument
•  
• IO Status in DCS
• CLP+ or CLP- means that the OUTPUT is clamped at a high/low limit in the system. This can be due to a lot of reasons such as the output to the control valve is full (control valve is asked to fully open) but the PV has still not reached the SV. In situation, the PID controller has failed. All other outputs from which became the inputs to this block will also be in a CLP status
• IOP+ or IOP- means that there is a high(>20mA) or low(<4mA) current at the INPUT. A low or high current can be due to a low of reasons such as transmitter is disconnected (open loop give 0 mA), transmitter set to burst due to sensor disconnection or any other electronic failure, and also transmitter PV reading is above URV or lower than LRV
• OOP data status occurs when there is no current to the output. This can be due to a discontinuity in the output or any other failure
• For Control valves, regardless the control valve is fail open or fail close, 0% seen in the DCS will always be close and 100% will always be open. Even though a fail open control valve requires 4 mA to to open it, the IO card will always invert the signal to 20 mA

Time Format

• Understanding Time Format
• The 24 hour clock goes from 0 – 24, the moment it reaches 24, it will be written as 0:00 with and a new date start her
• A new day starts at 0:00
• The AM/PM clock has no 0:00 PM or 0:00 AM. It starts from 1:00 AM to 11:59 AM. The next minute, it will become 12:00 PM. Then it will proceed to 11:59PM. In the next minute it will become 12:00 AM and a new date starts here
• A new day starts from 12:00 AM, 1AM, 2AM….. and finally to 11:59 PM
• 9/13/2009 0:00 is equal to 9/13/2009 12:00 AM.
• Time Standards
• The idea behind time standards is that a day should be equivalent to a full 360 degC rotation of the earth
• Anybody adopting a particular standard will all have the exact time at any point of time
• GMT is equivalent to UTC
• GMT AMT
• uses Canadian date format, mm/dd/yyyy hh:mm:ss AM/PM
• Is case sensitive
• No “Undo” button

Controller Performance Monitoring

• Is a practice to monitor controller performance and determine the root causes
• There are 3 main objectives of CPM, which is to report
• Controller Utilization
• Controller Performance
• Controller Diagnostic
• A CPM tool will typically analyse a sample of data. The Sampled data quality is typically measured and shown as the following KPIs
• Compression Factor
• A measure of the compression level in the tracking error (PV-SP).
• If the compression level in the given dataset is very high, the data cannot be trusted and may not be suitable for analysis. The Compression Factor is calculated as follows:
• Compression Factor = length of data / (length of data - length of compressed data)
• The length of data is the overall length of the entire data to be analysed, including all the interpolated points
• The Compressed data is the portion of data which has been compressed
• This is actually a poor way of describing it. Since length of data – compressed data = uncompressed data, CF is actually the ratio of length of data to uncompressed data. Uncompressed data in this case is the raw data obtained. So in can actually be rewritten as
• Compression Factor = length of data / raw data
• Compression factor larger than 3 is generally seen as no good, i.e. if the raw data contributes less than 1/3 of the overall data.
• We can also view compression factor is actually an expansion factor i.e. an expansion of the raw data
• Compression Factor is always known to cause problems. Commonly, the PA is not executed. One of the reason is due to the pins not correctly assigned as in PV assigned to SP and etc. Checked carefully!
• Controller Mode
• Auto – OP will change automatically to achieve SP
• Manual - OP changes manually, SP ignored
• Manual Tracking – OP changes manually, SP ignored, But SP will change to be equal to PV so that when controller changes from Manual to Auto, there will not be a bump in the process
• The KPIs are typically segregated by their analysis type
• Controller Behavior Analysis (Controller Utilization Analysis)
• Service Factor
• %Time controller is in auto or manual
• Output Saturation
• %Time the controller output is saturated (<0% or >100%)
• A good controller does not have output saturations
• Performance Analysis (Controller Performance Analysis)
• Performance analysis is an analysis done through use of process models which involves complex Fourier transform calculations. Since it is a ‘Performance’ analysis, It must benchmark against a good performance
• Benchmarking Parameters are typically set by the following parameters
• Benchmark response level
• The border definition for the settling time to be considered as stable in percentage. For example, if benchmark response level is 5%, this means that if the oscillation is less than 5%, the loop is considered settled
• Settling time
• Defined as the time it takes for the controller to stabilize at it’s set point < the benchmark response level in oscillations
• Autocorrelation (Displayed as a graph)
• Hardly used
• Determines how much a particular PV correlates with it’s past values
• A good controller will have less correlation
• Frequency Response (Displayed as a graph)
• Displays the response at different frequencies
• Typical analysis is to look at peaks which indicates oscillations
• Peaks at low frequency indicates detuned controller
• Peaks at high frequency indicated over tuned controller
• 
• This is a good controller as it matches the benchmark
• Impulse Response (Displayed as a graph)
• Displays how fast a controller reaches it settling time. It is actually the same thing as frequency response but instead of showing it in the frequency domain, it shows it in the time domain
• CPM software will need to model the process response to obtain this parameter
• Typical Analysis that can be done
• Measuring rise time
• Measuring settling time
• If the graph rises forever, or decays so slowly that it never reaches 0 – loop in manual, or controller saturated
• If the graph rises, then dying oscillations – usually an oscillatory disturbance. Possibly oscillations brought on by too much integral action
• Exponential decay from 1.0 to 0.0 – that’s what you want
• Graph Immediately drops from 1 to some value above 0, then slow dynamics
• in a slow loop – loop has significant measurement noise.
• in a fast loop – not enough control action
• Oscillations close to 0 in an otherwise fast loop – valve stiction
• Fast oscillations – over tuned
• Diagram below shows Matrikon CPM’s closed loop impulse response plot
• The red color is the benchmark impulse response, which a user defined in the benchmarking parameters
• In this example, the impulse response is bad. The settling time is 11 min in comparison to 2 mins of actual benchmark time.
• 
• Relative Performance Index (RPI)
• Is measured by taking the squared area under the curve of the closed loop response benchmark divide by the actual performance
• If the Actual performance is taking longer to settle than the benchmark, the value will be < 1, this indicates a poor performing controller
• If RPI = 1, this indicates a good control
• If RPI > 1, this indicates that the actual is better than the benchmark. This can typically mean that the benchmark time was set too loose.
• Economic Benefit Index
• The economic benefit measures how much potential benefit the controller has to realize if it’s working at the benchmark level. (In a way, it’s actually economic lost)
• It’s measured as
• If RPI < 1, EB = 1 – RPI
• If RPI > 1 ,EB = 1 – (1/RPI)
• One question one would ask, is how come if the RPI is higher, the EB gets higher as well? The reason is because EB sets a penalty for setting the benchmark time higher than expected.
• 
• Oscillation Analysis
• Is used to calculate Oscillation Index
• Oscillation Index
• A value between 0 to 1 which measures closer to 1 for higher oscillations
• Is calculated by performing autocorrelation by itself and taken as the amplitude of the first overshoot in impulse response
• Valve Analysis
• Is used to calculate the following
• Non-linearity Index
• Stiction
• Data Preparation
• One of the key thing in any condition monitoring tool is how data is processed before analyser
• Any condition monitoring tool will obtain a sample from a process, typically obtained from a historian or OPC server
• Reasons why data preparation is needed are as follows :-
• Missing or bad data
• Data is out of bounds
• Data needs to be linearized (e.g. using log linearization for pH or O2 analyser)
• Dynamic CPM tools such as Matrikon CPM will provide 2 expression fields for a user for each tags, they are
• The Evaluation Expression Field
• An expression executed on each points
• Will return either true or false
• The Replacement Expression Field
• An expression that executes when the evaluation field is returned as true
• The expressions are typically written in a programming style language. Predefined variables are used to refer to certain point of data. The table below is an example from Matrikon CPM. As can be seen, Matrikon uses last_good, point and next_good
•  
• CPM Programs will analyse controllers by their control scheme. Typical Control Schemes are:-
• Normal PID
• Cascade Control
• Selective Control
• No matter what the controller scheme is, A controller is a controller. It will always try to control PV at a SP by manipulating OP (or MV). The only thing that differs between controller schemes is the when and when not to analyse the control. For example,
• In a selective control scheme, we mat not want to run analysis on the controller if it is not selected.
• In a cascade control scheme
• In Matrikon CPM, when and when not to perform an analysis is manipulated in the controller mode.
• Ratio blocks for ratio controllers need not be analysed as it is just a simple multiplier. The controller is the one that can be analysed

Control Theory

• Control is the art of keeping a measured variable at a desired level
• Types of Controllers
• Open loop controller
• Feed forward controller
• Feed back controller
• The most common control algorithm used is the PID control. PID controllers are a feedback type controller
• PID controllers have 2 inputs and 1 ouptut: set variable(SV), process/controlled variable(PV) and manipulated variable (MV) (output)
• PV is usually taken from transmitters and MV is given out to control valves. The MV sent to control valves is a 4-20 mA signal. Setting either reverse or direction action in the controller will tell the controller the actuation direction to the control valve whether increase in mA will open or close the valve. For fail open CVs, giving 20mA will shut the valve instead of opening it.
• SV are usually set by panel man, some are taken from another controller’s MV (SV = MV) and some are taken from other subsystems such as APC (advance process control)
• The controller action will be to manipulate the MV till the PV = SV. This is a feedback system because the input is taken in front of the control valve
• PID control block can be set to normal mode, auto mode, manual mode or cascade mode.
• Usually, PID block have a set of tracking modes. Tracking in general means to force the output to a certain input. This is to ensure safe transition between manual to auto mode. The most common type of tracking modes are
• Manual Trackin
• Auto Tracking
• PID Controllers too has several types
• Simple Loop
• Cascade Loop – This controller is used if another controller (master) who also uses the same final control element gives the set point of the first controller (the slave). Reason why cascaded control is used is usually because the controlled process variable has a slow response. In this case, the master controller sending only set point values to the slave controller will control the controlled process variable. Example of an application is vessel level controls, where the level controller is the master sending out set points to the flowing out flow controller
• The term open loop and close loop is a state of the controller. For open loop, the input is not fed back, in this case the controller is not used, the input, which is the manipulated variable is manually set and the output flow (which would become the input to the PID controller ) will be observed. Close loops
• The concept of a PID controller is simple,
• There are 3 main components, which are :-
• The input, PV
• The set value, SV
• The controller output, MV
• The aim is to get the SV as close as possible to PV. This is done by creating an output of MV
• There are 3 elements in the PID controller
• The first element is P, proportional
• The second element is I, integral
• The third element is D, derivative

• The P Controller
• The P controller causes a change to the MV by an amount which is proportional to the change in Error or Delta Error. Error is simply calculated as PV – SV.
• The change in error is then multiplied by the proportional constant P. Some DCS vendors refer to P as proportional band(PB) , in which case P = 100/PB.
• The higher the change in error, the more the controller output will change
• The higher the P (Or lower the PB) the faster or more aggressive the controller works. However, this may introduce oscillations
• The problem with having P alone is that there will always be an offset between the settled PV value and SV. Why? It is due to the nature of the controller that it will try to get as close as possible to the set point value but no too close. If it does gets too close to it, it will force the reading to go further away from it. But if it gets too far it will come back closer. Sooner or later it will settle at a value which is not too close but not too far. Hence, the offset
• P only is hardly used due to this problem. It is only used for coarse control applications and cheap controllers like previously when they use to have pneumatic controllers
• The initial setting of PB is typically around
• For Liquid Flow/Pressure = 50-500
• PB should be high here because liquid response is typically very fast, so you don’t want to change the MV too fast as it will upset the process
• For Gas Pressure
• PB should be 1 to 50, i.e. the gain is high
• PB should be low here as gas moves and changes very slow.
• Please note that since P is always larger than 2x (i.e. 100/50 = 2), it means that the adjustment to the controller must be done higher
• For Liquid Level
• PB = 1 to 50
• PB should be low since liquid level moves slow
• For Temperature
• PB is typically around 2 to 100 (initial setting)
• PB should be low as D is very high
• For Analysers
• PB is 100 to 2000
• PB is low since we want a very slow controller response since analyzer readings are always lagging

• The I Controller
• To address the issue with the P controller, the I controller is produced.
• The I controller constantly adds a corrective action to the MV change. This corrective action is proportionate to the error (not change in error like P). The proportional constant is known as I (integral gain). Also known as repeat or Repeat/Min. It simply mean how many time the entire error is integrated into the controller within a 1 second
• Some DCS vendors also use Integral Time(T) instead of integral gain(I). Integral time is the inverse of integral gain. Integral Time = 1/Integral Gain. It means that how long does it takes for an error to be added into the system. It is also known as Reset or Min/Repeat
• To disable I,
• If the DCS defines it as Integral Gain/Rate/Repeat, Put I = 0
• Id the DCS defines it as Integral Time/Reset, Put TI = As high as possible! This is what we normally do when disabling I in a Yokogawa system, we put I as high as possible like 9999 or something…
• A high I or Low T means that the error will immediately be added to the corrective controller action. This is suitable for a fast moving control valve where it can react quickly.
• The problem with the I controller is that it is slow in action since the corrective action builds up slowly. We can add a high I (or small T) but this will result in oscillations. Why?
• Think about it. When we have I, we are actually adding up errors.
• The Typical I settings would be
• For flow control or liquid pressure, TI (integral time) is very small (<10s). This is because liquid changes it pressure or flowrate very fast, since it’s does not has an expansion factor
• For Gas pressure or liquid level, TI is normally a bit larger (30s). This is because gas pressure changes slowly. Again, this depends on the process but it’s better to start large and slowly go lower
• For temperature, also start at high TI, i.e. 30
• For quality, also start at very high TI, i.e. 60

• The D Controller
• To address the issue of the I controller acting very slow, the D controller is introduced.
• The D controller will add an acceleration of error (the rate of error change) to the controller movement. This acceleration is multiplied by a constant called the D or derivative time.
• By adding the acceleration, controller movement can be faster since the controller will calculate how much difference the rates are changing. If the change in error is constant, the D will have no effect
• D is usually used for temperature and pH control. This is because due to slow response. Even on cascade, the cascade loop will have a temperature controller
• Disadvantage of having D is that it will cause oscillation
• The typical D settings would be
• D is only used for temperature and quality, which typical values would be 20
• For gas pressure or liquid level, can use a bit of D but must be very small. Start at 0 D first and gradually increase

• Yokogawa Centus CS has the following controller function. Delta MVn is the change in MV
•  

• For P, MV will be adjusted by an amount equal to the rate of change of error. PB is the proportional band
• As you can see in the Yokogawa system, 100/PB encapsulates the entire equation. This is similar to a Gain (which is = 100/PB). The gain will encapsulate the entire equation
• Y_pid = Y_p + Y_i + Y
• _{= Kp ( e + 1/ Ti ò e dt + Td de/ dt )}
• For I, MV will be adjusted by an amount equal to the error it sel
• Delta T, is the control period, i.e. the cycle time of the controller. Typically it’s 1 second but there are some controllers that can be set to 1/3 of a secon
• TI is the integral time, in second
• As you can see, if TI = 5s, it will take 5 seconds for all the error to be summed u
• However, if this is not a typical controller with 1 second cycle time, delta T will be the alternative cycle time, hence athe TI will still maintain at 5 seconds even though the cycle time is not 1 secon
• Then again, you need to remember the proportional band thing. T or I will scale accordingly to PB or P

• Ziegler Nichols open loop tuning method
• allows to calculate PID parameters from the process parameters. The procedure:
• Step 1: Make an open loop plant test (e.g. a step test)
• Step 2: Determine the process parameters: Process gain, deadtime, time constant (see below: draw a tangent through the
  inflection point and measure L and T as shown. By the way: Today we have better and easier methods).
• Step 3: Calculate the parameters according to the following formulas:
• K = time constant / (process gain * deadtime)
• PI: Proportional gain = 0.9 * K, integral time = 3.3 * deadtime
• PID: Proportional gain = 1.2 * K, integral time = 2 * deadtime, derivative time = 0.5
  deadtime
•  
• Process gain = dPV / dOP, deadtime = L, time constant = T

• How do we know a controller has been tuned?
• ¼ decay
• Minimum Overshoot
• 

• Types of controllers
• Cascade Control
• Occurs when there are
• Multiple inputs or PVs
• 1 controller output
• The set point of the inner loop (Secondary loop) is controlled by the output of the outer loop (Primary loop).
• Operators will set the desired controller set point at the outer loop
• The inner loop is called the slave where the outer loop is called the master
• The purpose of having an inner loop is to reduce disturbances on the PV to be controlled
• Typically used for level and temperature control
• 

• Ratio Control
• Purpose of this controller is to set the ratio of flow between two parameters is constant even if one flow changes
• Only one of the flow will be controlled
• The ratio controller block will receive input only from the uncontrolled flow
• Typically used for blending
• 

• Override/Selective Control
• Is a controller which shares a same controller output for two controlled PV. At any point of time, the higher priority PV will be selected. The selection of which PV is more important will depend on an algorithm such as larger than or smaller than.
• Selective control is typically used in furnace FG control and steam heating control
•  

• Split Range Control
• Is a controller which has same PV but 2 controller output. The output is selected based on the range of the PV.

• Feedback Control
• A Control scheme trying to correct back something that has already happened in the past
• The disadvantage of using feedback control is that it will provide less stability
• Example of feedback control is furnace outlet temperature control

• Feedforward Control
• A control scheme which tries to prevent the error from happening in the first place
• A control is called feed forward when the following occurs
• The output goes to a calculation block which
• An Input taken from another instrument is fed to the calculation block. This variable is called the DV (Disturbance variable)
• Feedforward control should be used in any case except when there is process limitation

• Complex Loop
• 

Laplace Transform

• The idea of Laplace transform is to solve an equation containing differential and
  integral terms by transforming the equation into the S-space
• This is done by applying the Laplace transforming function as follows:-

{ f(t)} =

• The resulting expression is a function of s which we say F(s)
• Normally, people do not have to solve the entire formula as we can use standard
  Laplace transformations. Below is the table :-

 

Time Function f(t)    f(t) = -1{F(s)}	Laplace Transform of f(t) F(s) = { f(t)}
F1	s > 0
t (unit-ramp function)	s > 0
tn (n, a positive integer)	s > 0
eat	s > a
sin ωt	s > 0
cos ωt	s > 0
tng(t), for n = 1, 2, ...
t sin ωt	s > |ω|
t cos ωt	s > |ω|
g(at)	Scale property
eatg(t)	G(s − a)       Shift property
eattn, for n = 1, 2, ...	s > a
te-t	s > -1
1 − e-t/T	s > -1/T
eatsin ωt	s > a
eatcos ωt	s > a
u(t)	s > 0
u(t − a)	s > 0
u(t − a)g(t − a)	e-asG(s) Time-displacement theorem
g'(t)	sG(s) − g(0)
g''(t)	s2 • G(s) − s • g(0) − g'(0)
g(n)(t)	sn • G(s) − sn-1 • g(0) − sn-2 • g'(0) − ... − g(n-1)(0)

• The last 5 on the table shows the true power of Laplace transform. A complex
  integral calculation then would be impossible to be performed in the time
  domain, when transformed into the S domain can be easily represented
• Laplace transforms have several properties which are:-

·        Property1 - Constant Multiple

·        If
a is a constant and f(t) is a function of t, then L{a f(t)} = a L{f(t)}

·        Example
{7 sin t} = 7{sin t}

·        Property2 - Linearity Property

·        If
a and b are constants while f(t) and g(t) are functions of t, then

L{a f(t) + b g(t)} = a L{f(t)}
+ b L{g(t)}

·        Example
L{3t + 6t2 } = 3 L{t} + 6 L{t2}

·        Property3. Change of Scale Property

·        If
L{f(t)} = F(s) then L{f(at)} = 1/a F(s/a)

·        Example
L{F(5t) = (1/5)F(s/5)

·        Property4. Shifting Property (Shift Theorem)

·        L{exp(at)f(t)}
= F(s − a)

·        Example
L{exp(3t)f(t)} = F(s − 3)

·        Property5. Differential transformation

·        L{tf(t)}
= -F’(s)

·        Property6.

·        The Laplace transforms of the real (or imaginary) part of a complex function is
equal to the real (or imaginary) part of the transform of the complex function.

·        Let Re denote the real part of a complex function C(t) and Im denote the imaginary
part of C(t), then L{Re[C(t)]} = Re L{C(t)} and L{Im[C(t)]} = Im L{C(t)}