From conception through operation to decommissioning, the capital equipment lifecycle has been fragmented, with different people in charge of sequential stages. As the project and then the plant is handed off from one group to another, its information has been compartmentalized, kept in separate silos, and not in a single database or model.
Now we're seeing a shift in how companies want to deal with equipment information. This shift is supported by a convergence of capabilities in information technology and driven by an explosion in capital projects, for example, to produce and use hydrocarbon feedstocks from shale. In a market where, for many years, large projects were in the $50-100 million range, we're now seeing many $2-billion to $10-billion projects and some reaching $20-billion.
These projects are taxing the available resources of engineering, procurement and construction (EPC) companies, as well as owners, who are looking for better ways than Excel spreadsheets to control projects and bring them online faster and more economically. "We're limited in our ability to analyze data and make improvements when data is tied up in documents,” says Jennifer Garfield, engineering data project manager at ExxonMobil. "We waste effort turning data into documents.”
The opportunity extends into existing plants. "Brownfield information is incorrect, so we have to find, gather and validate it in the field—boots on the ground,” Garfield adds. "Incorrect data results in wasted capital, poor performance on start-up, schedule delays and safety issues. Higher efficiency is needed to handle increases in project volume.”
Just as automation professionals look to standards such as OPC and fieldbus, project managers and owners are urging development of standardized ways to pass data between design, engineering and construction applications, including electrical plans and P&IDs, and on to enterprise maintenance- and asset-management systems. Efforts such as Fiatech's (www.fiatech.org) ISO 15926 are extending standardization to allow one repository of information to represent a project from engineering and construction through operation, maintenance, migration and decommission — the entire asset lifecycle.
When you do your P&IDs, all of that information now can go into the engineering database, which can be rendered as two-dimensional views or drawings and incorporated in a three-dimensional model. This master data model can be used to feed asset dashboards, representing the as-built and as-is plant or facility throughout its life.
Having a master data model, one version of the truth, allows the formerly separate and sequential build, commission, turnover, operate and maintain steps to interact, and even be done simultaneously, eliminating surprises and improving results. It's also important to process safety management, where using incorrect data has detrimental effects, and causes plants to show up in the evening news.
Simulations using the 3D models also are especially useful for training operators and maintenance personnel. "The brain is good at relationships, not details,” says Dave Wheeldon, chief technical officer and head of engineering and design systems at Aveva (www.aveva.com). "It works best in 3D because that—not Excel spreadsheets—is how we avoided saber-toothed tigers.”
Call It BIM 1, 2, 3
Originally standing for building information management, the acronym BIM is now used to broadly refer to systems for managing data associated with complex assets. Bentley Systems (www.bentley.com) identifies three levels of sophistication in BIM (Figure 1). Level 1 uses 3D CAD models for design and engineering, and is well established within the EPC community. After design, the virtual model is typically converted to conventional 2D drawings, lists and schedules for construction.
BIM Level 2 expands the level of "optioneering” — the ability to explore alternatives — across multiple disciplines (including operations and maintenance) through simulation and analytical software. It extends the live data model into construction to accommodate changes as problems arise and designs are revised. Level 2 implementations are now expanding, driven in part by a mandate by the British government to use it in all public projects.
Bentley envisions BIM Level 3 as extending the lifecycle of the model by making as-built and as-maintained models available to operations and maintenance, while "big data” from sensors and operating metrics contribute to a "rich, immersive” information model to improve performance, safety and sustainability.
Until recently, "Ninety percent of [BIM] activity was with EPCs,” says Frank Joop, executive director, business development, Intergraph (www.intergraph.com). "Owners waited for the project to be finished, then got a trainload of documents, half of which were obsolete. Now owners want things differently. They recognize not only the cost to build, but also the cost to maintain. So they want information for the operational stage.
"We need to change from CAD — drawings, lists and reports — to a smart plant environment centered around data,” says Joop. "Drawings and reports become a byproduct, views of the data, but not the data itself.”
Access Aids Collaboration
Traditional project design and engineering has been sequential, for example, chemical, then electrical, then mechanical. "Now all disciplines can work in the same space, on their own part of the project, but in context, through their own diagrams or portals,” says Wheeldon. "Instead of a small picture, they can work on part of the big picture—connected systems, say, fluid and control, on a living asset, changing it before it's built, with multiple simultaneous sub-projects, and all the while, they're guaranteed to be seeing the absolute truth.”
Rules-based engineering can be programmed with constraints, easing decisions and preventing errors. "If you have a pipeline and a valve, the valve knows it's on the line and where it is,” Joop says. "Rules check that the valve is the right size and in spec for the line size and process conditions. They know if a ball valve should be a gate valve.”
The 3D model can be reviewed with construction, operations and maintenance personnel. Critigen (www.critigen.com), a geospatial IT services company, does this using a computer-assisted virtual environment (CAVE) — a meeting room with walls of screens that allow participants to walk through a design at any scale needed to go over projects, identify problem areas and collaborate on improvements.
When the model is kept current through operations and maintenance, it can be used to manage programs such as leak detection and repair or calibration. It can be interfaced with a computerized maintenance management system to schedule maintenance while storing and providing documents and data.
"It's also a path to rigor through traceability and records of installed components,” says Wheeldon. "If you find that a component has a reliability issue, say, an orifice plate is wearing, you can find comparable devices in similar applications by querying the rest of the assets. Knowledge-sharing is no longer confined to one facility or one discipline.”
Methods to render existing plants in 3D have come a long way. "Hexagon uses Leica laser-scanning to build a 3D model of ‘cloud points' — millions of dots that can be assigned attributes,” says Joop. "These are overlaid with high-definition pictures so when you select a perspective, the software knows what you're looking at. When you click on a tag, it goes into the software to access all the data related to that tag. Or you can look for a tag, and it shows you where it is.”
The models also can be used in 3D simulators for training, with avatars controlled by operators, to plan work or to practice emergency and safety procedures. "The gaming generation has the expectation that they can test themselves, develop a mental model, learn the risks and dangers,” says Wheeldon. "We can tap into that gamification mindset of experimentation, of what happens when you open that door.”
Implementation: A Matter of Give and Take
These amazing software system capabilities are up against a significant headwind of their own limitations as well established practices, mindsets and turf-protection in established plants and companies.
"At handover, owner/operators (O/O) currently receive an avalanche of documents and some data. The O/O enters some data into operations and maintenance (O&M) tools,” says ExxonMobil's Garfield. "Some documents are not received, and some are dumbed-down. The EPCs say O/Os must tell them what they want, and the O/Os don't know what they don't want. They need to be educated and to have internal standards.”
O/Os don't want to dictate what tools the EPC must use, because it increases the cost. "But they can dictate the characteristics of the data—open vs. closed, and compatible with their systems,” Garfield says. "They may have to pay 1% to 2% more, but it's worth it to avoid the extra work and potential litigation.”
Southern Co. (www.southerncompany.com) is requiring EPCs to provide data as well as documents. Until they deliver the data model, they haven't fulfilled the deliverable and earned the credit.
In the plant, maintaining data integrity can be challenging. "We need to delete 2D drawings the same way we destroyed paper copies,” says Barbara Migl, CAE technology leader, Dow Chemical. "The 3D model is the master document: What needs to be updated before the job is done, what everyone accepts as the master. At Dow, we divided electronic and paper drawings. Some systems are on one; some ore on the other, but none are on both. 3D CAD is the master for Dow-created and maintained projects. 2D drawings are tossed and recreated when needed.”
But for supplier projects, unmaintained 3D models are tossed, and 2D drawings are maintained as masters. "But maintenance would not update the 3D model. They didn't think it was their place to do engineering work, so they made new measurements and their own 2D drawings,” Migl says. Dow is stopping that by giving them people who can access and update engineering files.
"But the instrument group keeps the instrument database as its own master document,” Migl says. "Their updates are processed separately into the 3D model. So the proposal is to put it all into one data warehouse. Maybe we will, but then, where is the master data?”
The problem of maintaining the master also comes up during modifications. The plant needs an accurate as-is, and some planned changes are never made. "Once you have the data-based model, you have a much better chance of keeping it current,” says Joop. "It's the single source of the truth, not a collection of sheets here and there.”
With an accurate as-built, when you need to consider a change you can carve out a section of the plant and set it outside in a project (Figure 2). "The plant keeps going and if another change is made, say, if a spare terminal is no longer spare, you get a notification,” Joop says. Use regular design software, do what you need to do, and when it's done, copy it back into the as-built. If the project is cancelled, the as-built is not affected, so it's not a problem.
In as-built systems, data integrity from document to operations and maintenance can seem impossible, but it's worth it to try to do an imperfect job. "Don't let perfection be the enemy of good,” says Will Goetz, CMRP, CMO, Management Resources Group (MRG, www.mrgsolutions.com). "Build around tags and asset numbers, automate what you can and enrich the database by secondary processes. Walk down what you must.”
BIM Payback Centers on Efficiency
Like most information technology, return on investment (ROI) on BIM tends to ignore the indefinite possibilities and focuses on automation of existing workflows.
Kenneth Barry, senior technical leader, Electric Power Research Institute (EPRI), recently reported on EPRI efforts to estimate ROI for data-centric configuration and asset management. EPRI surveyed member companies to determine the hard and soft costs of integrating a single database from different starting points, as well as the savings realized in engineering, licensing, procurement, operations and maintenance. Return on investment is in the form of time savings to find and verify data, in workflow and through electronic signatures, and through collaboration between design and operations.
A detailed study of personnel and time requirements in operating plants is to be published at the end of the year. Preliminary results vary, with higher returns on more data-intensive activities, but show typical project paybacks of three to eight years, Barry says. "These are on existing plants that are expected to last for 80 or 90 years.”
ARC estimates inefficiencies that can be eliminated by BIM add up to about 1.5% of total revenues. "At Dow, that's $900 million,” says Dow's Migl. "I started collecting data and not just ballpark estimates, but examples of people hired just to do data translation and find stuff, to print out and type in. I found $36 million, and that's not all of it. That's a lot of money we're wasting.”
In another example, Dow wrote a tool to check consistency across four instrumentation P&ID systems using ISO 15926 as a standard, "Just to check the instrumentation specifications,” Migl says. "The first project saved half a million dollars just by rationalizing instrumentation attributes.”
The savings by managing data more efficiently may be easiest to quantify, but the potential for improving operations, maintenance and safety is the most exciting. Plants that don't include input from operations at the design and engineering stages will see 9% higher capital expenditures, longer and more costly ramp-ups, and higher expenses to operate and maintain, says Goetz. "You may never get back.”
The Motiva refinery in Texas, which is so old, it was originally fitted with wooden pipes, decided to capture all its existing data to reduce operational costs and risks, says Joop. "Now, if they see a gasket leaking on an old valve, in minutes they can know exactly what it is and what they need to fix it.”
Master data is increasingly important for efficient regulatory compliance. For example, "API 691 says there are certain assets you have to build in a maintenance and reliability strategy for,” Goetz adds. "Without robust asset master data, you will have a house of cards.”