Key points
First some definitions:
NOA – Namur Open Architecture – invented by the User Association of Automation Technology (NAMUR) this is a way of getting additional values and asset management data (sometimes called stranded data) from an instrument by installing a second parallel standard communication channel which could be wireless, Bluetooth, 4G/5G etc. and does not affect the existing primary automation system connection for the measured value.
Ethernet – APL™ – Ethernet Advanced Physical Layer – was created by a trade body of equipment manufacturers including OPC and will provide Ethernet to the field, to individual instruments and sensors. Still under development it will enable long cable lengths and explosion protection via intrinsic safety with communication and power over two wires.
OPC UA – Open Platform Communications Unified Architecture – is a machine-to-machine communication protocol for industrial automation developed by the OPC Foundation a group of equipment manufacturers.
MTP – Module Type Packages – is a developing standardised methodology, again from NAMUR, that provides a common language for describing the properties of process modules independent of manufacturer and technology. This makes it possible for any higher-level automation system to use MTP to accurately control a specific Module/Package.
Process Automation I/O System Market Size
The overall market size for process I/O systems is estimated at just over $3bn, growing at around 3.7% CAGR. Siemens have approximately 14% of this sector with ABB at 10%, Honeywell 9% and Rockwell 8%.
General Trends in Process Automation I/O Systems
- Decentralisation – distributed installation of remote I/O stations in the field instead of centrally in the main control room
=> reduction in plant cabling
- Decline in plant Fieldbuses
- Wiring of the field sensors directly to I/O terminal boards
=> elimination of cross-wiring cabinets
- Greater flexibility with Universal I/O aka Late Binding I/O aka e-Marshalling aka Configurable I/O
=> decouple automation hardware and software, shorter system deliveries (time to market) and earlier field wiring
- Increased implementation of “intelligent” sensors with higher data set
=> advanced control optimisation, remote diagnostics, and asset management data
- The rise of Ethernet based communications into the field
=> e.g. PROFINET and Ethernet – APL™
A LinkedIn straw poll of the Automation community produced the following results:
It is no surprise that “Remote I/O stations in Field” won easily in all 3 groups but what is interesting is that “Type of Universal I/O” was beaten into last place in the Instrument Engineers and ISA groups which are considered more process automation biased than the factory automation focussed Automation group. “Plant Fieldbus” also did better than expected taking second place in all groups.
Decline in Plant Fieldbuses
There has been a historical trend over many years towards decentralisation in the field instead of back to the main control room. The most extreme example of this was the use of plant fieldbuses to save on plant cabling. Paradoxically there has been a decline in plant fieldbuses whereas the trend to use remote I/O stations even into hazardous areas has continued.
There was a time when fieldbuses were all the rage particularly in the oil & gas and petrochemical sector. This has changed!! Anecdotally a major oil company explained that 20 years ago an engineer had to justify why they were not using Foundation Fieldbus (FF) for a new project, 10 years ago this changed to no preference until now the engineer must justify why they need to use FF rather than a conventional 4-20mA HART solution.
One of the main perceived benefits of fieldbus was cost saving and simplification of plant cabling – this saving is greatly reduced for hazardous area deployments and along with the level of technical competence required to design, install and maintain fieldbus systems has caused the massive drop in favour. FF installations allegedly had numerous earthing and grounding issues. Many of the remote diagnostics and asset management benefits of fieldbuses are realised with conventional HART over 4-20mA signals anyway, and indeed the second parallel channel proposed by NOA provides this without affecting the main automation system connection
Wireless which offers even more benefits for plant cabling is growing in popularity and taking some of the FF market share but mainly for indication only, low scan rate, non-control measurements. However, with wireless not only transmission routes should be considered, but also the fact that if the transmitter is not self-powered and it requires power cabling anyway – is this really a saving?
Why has PROFIBUS fared better as a Fieldbus?
PROFIBUS has fared better partly because it is established in industry sectors outside oil and gas, has become the electrical system standard for MCCs and VSDs to communicate and is used by many vendors as their I/O bus. PROFIBUS is also evolving into PROFINET a super deterministic protocol with ethernet speeds, media and components. PROFINET is especially suited for remote I/O communications as used by Siemens.
Why the Perceived Popularity of Universal I/O?
Again, this is counter intuitive against the trend for decentralisation, back to universal I/O in the main control room. But certainly, this is happening, driven by the DCS manufacturers with the plethora of universal and configurable I/O systems being released in the last few years. However maybe the earlier polls show this is not so popular with the end user automation engineers as the DCS companies would have us believe.
Universal I/O takes many forms, some with 16 or 32 channel I/O modules which are configurable per signal, others where the required module is selected on a per channel basis. The benefits are reducing automation system delivery times, early field wiring installation; and handling I/O changes very late in the project lifecycle.
But many types of I/O are required – powered and unpowered, 2 and 4 wire, simplex and redundant, safe and hazardous area, non-SIL and SIL, increasingly short-circuit and open circuit protection, galvanic isolation etc. – not all universal I/O systems can handle all these yet.
For example, if a solution requires separate intrinsic safety barriers this rather defeats the object. Universal barriers are now available but more expensive than an I/O card with integrated barriers of course. In addition, many universal I/O systems do not provide true I/O redundancy.
It is a myth that universal I/O saves space by eliminating marshalling and cross wiring cabinets – this is more a product of the trend to terminate field cabling directly on the I/O board (or more accurately to a fixed termination unit so that the I/O board can easily be removed). Indeed, universal I/O is considerably less dense than fixed I/O in the range of 55 – 140 I/O per m compared to 400 – 500 I/O per m, so needs much more space than fixed I/O.
Is Universal I/O more cost effective?
Surely a dedicated fixed, high density I/O module is more cost effective than a universal solution which must have every I/O type built-in or selected on a per channel basis? Add to this the cost saving for plant cabling when dedicated I/O is used remotely close to the process unit and the reduced footprint – this must be a winner!!
Of course, a system that combines high density dedicated I/O with the late binding capabilities of universal I/O is also a strong argument. This can be achieved when multi-channel modules are plugged in to a generic backplane at the last minute without affecting the already terminated plant cabling.
For example, the Siemens ET200SP HA I/O can handle up to 896 signals in one standard cabinet and the units are modular so that the required I/O and termination type can be plugged in to a universal backplane when required. Add to this using the configurable card (16 signal selectable AI/DI/DO per channel) for the last 10% in case of changes is perhaps the ultimate panacea for late binding, cost effectiveness and space.
Example Direct Connection Cabinet 56 I/O Modules, up to 896 I/O 800 x 2000 mm (width x height)
Features:
- One side access, 400mm deep
- Cabinet ambient temperature -20..+40°C
- 120/230V AC redundant power infeed
- direct connection up to 2.5mm²
- PROFINET communication via copper or fiber optic cable
- Up to 56 field cable (20mm diameter)
1 - Infeed MCB and RCB
2 - AC/DC SITOP PSU 8200 40A(Siemens, 6EP3334-8SB00-0AY0)
3 - Redundancy module SITOP PSE202U (Siemens, 6EP1961-3BA21)
4 - Power distribution and cartridge fusing
5 - add. HW, e.g. SCALANCE XC206-2-SFP (Siemens, 6GK5 206-2Bs00-2AC2)
6 - T 200SP HA with interface module and bus adapter
7 - Spare wire and M (ground) terminals
8 - Profile for cable fixation and shielding
9 - cable duct
Engineering Contractors (EPCs) favour Late Binding I/O!
EPCs prefer late binding because they can specify the I/O very late in the project lifecycle. It has long been known that instrumentation is finalised after the piping, mechanical and process design are completed, and any change after that eventually filters through to the Automation System.
Universal I/O is not the whole story of course, the DCS must be capable of being configured so that the software addresses I/O points in a generic way and not specifically rack-module-channel. This allows the application software to be decoupled from the hardware build and developed early in the project before the I/O is finally defined.
Ethernet into the Field
The trend towards smarter instruments/sensors has led to the initiative of Ethernet into the field – first as the I/O bus to remote I/O stations in the field – originally PROFIBUS DP then PROFINET is widely used for this now. Then eventually to Ethernet directly to the individual instrument/sensors. This trend will surely weaken the argument for Universal I/O.
A Truly Universal Plug and Produce Solution
An example of a truly universal plug and produce solution is the Siemens Compact Field Unit (CFU) utilising PROFINET (soon to be Ethernet – APL™ versions) to distribute the unit remotely close to the process and then up to 8 PROFIBUS PA instruments can be connected and automatically recognised on any port. This provides flexibility, simplification and massive savings in plant cabling and junction boxes etc. This unit fits well with the Ethernet into the field trend and could produce a renaissance for Profibus PA instruments.
From Instruments to Packages to MTP
The trend towards smarter instruments could be viewed as instruments becoming closer to packages or modules, in their own right. Indeed, NOA and MTP, both developed by NAMUR, are complementary solutions. MTP more important for green field installations whereas NOA comes into its own for brown field existing sites. When NOA is used for green field applications ideally it would use OPC UA for communication.
MTP Modular Automation and NOA NAMUR Open Architecture are Complementary
Conclusions
- There is a historic trend towards decentralisation of I/O stations however the decline in fieldbuses and the rise of universal I/O are both bucking this trend. Whereas the continued growth in remote I/O and the ethernet into the field initiative prove this decentralisation trend is still valid.
- Fieldbuses especially Foundation Fieldbus are declining because of technical competence and historic problem installations.
- PROFIBUS is still popular because it is used outside of oil & gas, has become a defacto standard for communications in electrical systems and many vendors use PROFIBUS DP and now PROFINET as an I/O bus.
- EPCs favour late binding I/O solutions because they can define the I/O at the last minute in the project lifecycle, automation system deliveries are reduced, and plant cabling can be installed and terminated earlier.
- High density dedicated I/O solutions can be more cost effective and space efficient especially when used remotely to save on plant cabling. Dedicated fixed I/O systems have a density as high as circa 470 I/O per m whereas universal I/O can be as low as around 55 I/O per m.
- In practice it is likely that fixed, high-density remote I/O systems, universal I/O, NOA, and Ethernet – APL™ will all co-exist for some time becoming preferred in certain areas. But will any one strategy prevail? Only time will tell – stick around for an interesting ride!!
