Record collection

As the data generated by typical industrial processes grows apace, sectors as varied as water, oil and pharmaceuticals are having to look at how they collect and use it. Colin Carter explains.


Data acquisition has been an issue since the first sensor was connected to a system. Eliminating man from reading dials or displays and writing the results in a log book represented a step forward in precision and reliability, and it is now the norm for machines to send information to machines with no direct human interaction.

The amount of data generated by typical industrial processes is growing rapidly – a direct result of the increasing use of smart sensors, enhanced predictive maintenance regimes and better condition monitoring of plant components. Plant intelligence is growing and Supervisory Control and Data Acquisition (SCADA) systems are taking advantage of this.

The US-based ARC Advisory Group estimates that the worldwide SCADA systems market in the electric power industry will grow at a compound annual rate of around four per cent until 2009, driven partly by the ease with which such systems can be integrated into enterprise systems.

In the oil, gas, water and wastewater industries the picture is similar, with the group estimating the world market will be worth close to $900m by the end of 2009. One driving force for growth is legislative: more monitoring and higher data integrity is ever more a legal requirement, and these sectors are having to look at how they collect data and what they do with it.

Condition monitoring and predictive maintenance is one way data is used intelligently. Alan Davies of Machine Condition Monitoring explained. ‘Condition monitoring or CBM (Condition Based Monitoring) is an effective form of predictive maintenance, forming an integral part of the modern approach to maintenance in capital-intensive process industries,’ he said.

‘Condition monitoring involves studying some aspect of a plant during operation to detect a condition that might lead to premature failure, and then determine the cause of that condition.’

For large-scale plant there could potentially be hundreds of components for which failure means downtime. And in industries such as oil and gas this can be very expensive.

Rahul Kulkarni, product manager for Industrial Data Acquisition and Control at National Instruments highlighted some of the challenges. ‘For applications such as accelerometers used to analyse vibration, much data is needed – for the frequency response analysis 24-bit resolution is required and most programmable logic controllers (PLCs) are 12 or 16-bit maximum. The processing for such applications is best carried out by PC rather than in the field – using something like National Instruments’ Labview virtual instrumentation.’

Reliability of such condition monitoring systems has also been assisted by the internet to relay collected data. Monitoring can now be carried out remotely from anywhere in the world. Some solutions (such as Machine Condition Monitoring’s t-mac digital dashboard) use wireless connections and GPRS to communicate plant data to servers anywhere in the world. Alerts can be also be sent via e-mail and SMS.

Use of the web is growing – although as far back as 1999 Telemecanique produced a web-enabled PLC, with an internal web server to enable data to be retrieved via a normal browser interface.

Dave Sutton, Schneider’s product manager, HMI & PLC software explained: ‘Being web-enabled devices, PLCs can hold comprehensive internet sites of their own so both real-time process data and diagnostics can easily be viewed at source, as opposed to having to be stored within traditionally expensive and difficult-to-configure/maintain middleware (SCADA).

‘Latest generation PLCs (such as Modicon Premium with FactoryCast Web Server) are now supporting web services, which means that business/reporting software can retrieve process data direct.’

Recent large-scale installations of data acquisition systems include the Habshan gas plant expansion project in the United Arab Emirates, where an Invensys Scada system will monitor and control gas wells and remote manifold stations.

The system, which is built around Wonderware, employs extensive redundancy on all production-critical components – from intersystem communications modules to Foxboro-based remote terminal units (RTUs). The RTUs will be installed at over 180 sites in the complex, at gas production wells, remote manifold stations and gas re-injection wells, and they contain over 10,000 I/O points.

Field RTUs will send information via a dedicated two-channel UHF radio data system to the control room. They can be polled over four possible pathways, ensuring data is always retrieved by the control room servers – which themselves are connected via a high-speed Ethernet LAN, again with built-in redundancy.

The water industry also has applications requiring the acquisition of much data. In the US, the District of Columbia Water and Sewer Authority (DCWASA) has installed Emerson systems to control operations at its Blue Plains treatment facility and to monitor its water distribution and wastewater collection system.

Employing Emerson’s Ovation control architecture, the systems manage the entire water/wastewater network through an integrated process control system incorporating redundant configurations.

Technology for automatic monitoring and control of primary sedimentation basins and pumping systems will be integrated, with operational control systems for primary/secondary treatment, filtration and solids handling. It is estimated that the plant will have over 40,000 I/O points – with something like 60 RTUs, 25 redundant controllers and one central control facility.

What is claimed to be one of the world’s largest SCADA systems has been installed at Singapore‘s Changi water reclamation plant. configured with half a million tags distributed over 50 nodes, the plant uses GE Fanuc’s iFIX SCADA application to collect data from functions such as tunnel pumping stations and processes such as solids removal.

The Changi system uses real-time portal and terminal services all accessible via standard web browsers. This information is fully integrated into the plant’s business system, and can be collected from anywhere in the plant.

On a smaller scale, data acquisition has been a major issue in the processing of blood plasma at the Bio Products Laboratory (BPL) plant at Elstree in Hertfordshire. This process has to conform to legislative requirements for the validation of process data, for both regulators and manufacturers.

As part of the separation process (for plasma and protein fractions) the blood is centrifuged. The process parameters are critical and the results if the process fails are the loss of a whole batch of 6,000 litres of plasma.

Throughout the process the (sub-zero) temperature of the plasma must be controlled to within 1oC. A Schneider PLC-based control system takes temperature, flow, pressure and level data from process sensors and feeds the data into a system, which complies with FDA 21 CFR 11 legislation on traceable records and signatures for verification.

The whole process can be monitored from a control room or from anywhere on the plant – for the hazardous area I/O and terminals MTL hardware was installed.

As for the future, Kulkarni highlighted a couple of trends. The need for intelligent data acquisition will grow, so sensors will be required to convey information on how variables are changing rather than fixed data so that trend analysis will be the norm.

He also sees a blurring of boundaries between smart devices, control systems, SCADA systems and enterprise systems, with shared platforms and open connectivity the norm. But whatever happens, there will be more data extracted from systems and the challenge will be to use it ever more intelligently.