The OSIsoft User Group held in Amsterdam last month had over 650 attendees, including 100 OSIsoft personnel. OSIsoft founder and CEO Pat Kennedy’s keynote described the ‘meltdown economics’ of the current global situation. Oil price uncertainty means that projects will be canceled or scaled back and that ‘cash is king.’ There will be a slow down in projects and a shift to ‘low hanging fruit’ in the form of small projects. Companies will seek to leverage information. Small projects can accumulate to interesting result. Back in the last downturn, in 1982, OSIsoft noticed that sales increased as companies sought to optimize their business in challenging times. More recently, one client combined hundreds of mini projects targeting energy savings which led to a $30 million per year saving. The same kind of approach should help companies plan for an oil price in the $50-$200/barrel range and make sure their projects work ‘in all scenarios.’
While process control is a very broad church, spanning continuous, batch and discreet/manufacturing activities, OSIsoft’s PI System, with its focus on real time data capture has a particular affinity with both upstream and downstream oil and gas operations. It may be something of a paradox for the IT professional, that while the PI System is at the heart of process monitoring, it has little to do with process control. By way of an explanation, we start this report with a paper that presents the situation at the ‘coal face,’ the interface between PI and process...
Maria Aniorte of the Spanish utility Iberdrola outlined the problem of optimizing a complex plant with some 10,000 PI System tags and hundreds of process control loops. If control loops are not working properly the result is poor plant performance and oscillations in output. But optimizing across hundreds of process control loops and coupled valves can get very complicated—requiring ‘deep math.’ Iberdrola’s market analysis of optimization tools resulted in the acquisition of Expertune’s Plant Triage (PT). Iberdrola now monitors around 600 control loops through a central PT server that rolls up information into a summary dashboard showing plant performance. Insights from PT have allowed Iberdrola engage equipment vendors to adjust things like pump parameters. Savings can be ‘high to immeasurable’ if accidents are avoided. Asked if there was bi directional data flow to the plant Aniorte replied no—‘We make recommendations to the plant operators. The people in the plant take action and make the changes. There is no Plant Triage output to the DCS.’
OSIsoft’s Deter Van Wordragen described how PI System is being used for power monitoring in the data center. Flagship client Microsoft is currently spending $4 billion on four new data centers. One, located in Chicago will run MSN/Hotmail/Virtual Earth and is to consume some 60MW of power. Microsoft’s servers are managed with PI for power efficiency—one data center has around 600,000 PI tags.
In a panel session, Dick Wernsing outlined how PSE&G is extending its maintenance cycle by adjusting the maintenance interval to operational requirements—in a shift from ‘emotional’ decision making to data based decision making. This has been achieved by a link from the SAP asset registry to the PI historian to produce ‘just in time’ work orders. PI gives visibility into what’s happening in the plant so that a minor problem is fixed with an inexpensive repair before it results in a costly failure. It is also important to make sure that information gets noticed. So that for instance if a compressor runs over 2 hours/week (indicative of a leak), it gets attended to. ‘Don’t let the engineers get in the way, they make rules but should not be able to change them every time they get a notification.’
John de Koning described how Shell Chemicals is standardizing its multiple equipment and laying down the foundation of an operational data infrastructure. The key is a standard data acquisition layer. But for Shell, PI System ‘standardization’ is about more than just installation, it involves standard processes and design. Often PI is deployed to address a particular technical issue. But how do you justify continued deployment once a problem is fixed? It is rather like the justification for a road—you have to have a foundation to build on. Equipment is getting more intelligent. Rotating equipment and high pressure pumps deliver huge amounts of data. It required a significant investment in a brownfield site to get all this into the PI System. The situation is easier for greenfield sites. Shell’s PI systems hold 400,000 points which leads to data quality assurance issues—‘the more data, the harder it is to assure quality.’
Sarah Al-Aqaily told how PI is the foundation of the integrated control and information management (ICIM) system for RasGas’ North Field. RasGas is the largest LNG producer and transporter in the world with exports to Japan, Taiwan, the US India and the EU. By 2010, capacity is planned to reach 37 million tonnes/year from seven gas trains and 2 helium recovery plants. The ICIM vision is to expand the scope of automation and information management across all of RasGas’ activities. The system assures data integrity and avoids data duplication. Information flow is outbound to other systems—the ICIM can’t write or control the plant. Real time data from the process, laboratory information management systems, hydrocarbon allocation, revenue and SAP/ERP for operations, commerce, shipping and finance are all collected into a real time information system (RTIS) which assures the accuracy of metering data that is passed on to financial systems.
Christian Roller traced Microsoft’s history of involvement in automation dating back to the Windows-based HMI (1985) and COM/OPC (1990). Today, Microsoft supports a ‘rich’ vendor ecosystem including OSIsoft, AspenTech, Siemens, Emerson and many others. Pasha Ahmed (OSIsoft) joined Roller to sketch out how Microsoft technology was deployed by Saudi Aramco. Aramco addressed the challenges of scalability and data validity by implementing PI System and Microsoft tools—notably SharePoint. Aramco appears to hold the record for tag count at around 2 million all funneled into the control room for display on a huge 70x3 meter screen.
Richard Coomber presented Shell’s Production Portal Architecture. This leverages OSIsoft’s latest AF 2.0 technology. PI AF provides a consistent representation of a plant for analysis with tools such as ProcessBook and Excel. The AF SDK is used to tailor a system to an organization’s specific requirements. Shell found that the move to AF revealed considerable data quality issues with ‘non standard’ tag names. Data quality management is now plumbed into the network—notably using the µ-QA tool from Magion. Process Book provides a geographical view of assets with drill down to offshore meter data and status. The Production Portal is built atop SharePoint, PI RtWebMarts and ESRI ArcGIS. µ-QA checks that meters are calibrated, that the network is up and the tag data is good. When a problem is detected, an SAP workflow is triggered.
Nancy Shifflet described an interesting use of PI when a NiSource compressor station got hit by a tornado. Using Intellution’s Ifix on a PC and an interface to PI, NiSource got the station back in operations in a couple of days despite limited communications and power. A late as 1998 NiSource had mostly paper-based SCADA logs and physical archives. Since then a PI System has been deployed with ‘human’ tag names representing location and facility names. Today PI is used throughout the company—by HSE, facility planning, field services and marketing. Applications include regulatory, asset performance, lost and unaccounted for gas and fuel consumption. PI scope has expanded from gas control SCADA to embrace electronic measurement, weather data (for planning and operational support)—and modeling. AF has brought a global data view to optimization analysts. FERC, State and Federal and environment reporting is all done instantly from PI. ‘PI provides connectivity, performance, scalability and ease of use to gas controllers, engineering and field services and environmental groups.’ In the Q&A, Shifflet was asked how much value NiSource put on its real time infrastructure over the previous SCADA system. She replied, ‘If I took PI away, users would kill me! It is so embedded. Otherwise we now know much more about our capacity although it’s hard to put a dollar value on this. PI is used from execs to roustabouts.’
OSIsoft VP R&D Ray Verhoeff took a peek at the future of standards, in particular, the soon to be released OPC Unified Architecture (UA). This is to expose a new address space, OPC objects and relations from an OPC Server. Mapping has been achieved between AF and UA. The UA architecture ‘promotes’ an information model—exposing nodes and relationships such as wells, tubing node, casing node as a hierarchy. Different industry groups need to standardize definitions and then ‘the software will comply.’ Verhoeff offered a long list of standards—he has been personally involved with PRODML. While this group is not yet ready to endorse UA, it is recognized as a ‘strong offering’ and the situation may change in a year or so. UA addresses security issues. UA transport profiles are no longer COM based and so allow for non-Windows platforms through XML-based Web Services. UA XML is defined as an XML Schema plus WSDL running on HTTP/HTTPS or UA binary. PI System can act as an OPC UA server to feed Windows, Linux or embedded OS clients. OSIsoft’s new architecture has been released as V1.0 with a PI JDBC Bridge from Linux (Ubuntu 7.10) talking to a PI OLEDB (Windows Foundation Class RDS) on the server.
Michael Lamb, (Xcel Energy) believes it is ‘time ripe to blow up Edison’s electricity generation utility,’ to move from load following to load shaping. Xcel is the US’ N° 1 wind provider, with 3GW installed capacity. Smart Grid City (SGC), a project to create the ‘utility of the future’ in Boulder, CO, has 3.3 mm residential customers with currently one meter read/month. But there could be 20-30 readings every 5 minutes. A typical Smart Grid activity could involve a plug in Prius used as a 10 kWh energy source used over night for load shaping.
Finally our salesperson of the show award goes to Transpara’s indefatigable Mike Saucier who manned the coffee machine, monitoring who was drinking what and displaying a running total on his Blackberry with Transpara’s Visual KPI!
This report is an extract from The Data Room’s Technology Watch report of the OSIsoft user group—more from tw@oilit.com.
© Oil IT Journal - all rights reserved.