As offshore platforms are automated from top to bottom, and integrators stitch together disparate systems and protocols into a smoothly-running continuum, the inevitable next step will be assimilation of the performance data with onshore management systems.
Because management execution systems (MES) and enterprise resource planning (ERP) continue to mature and evolve right along with technology and business practices, it appears likely that offshore drilling and production platforms, with their sophisticated technology and vital linkages, will become the launching pad and proving ground for a new generation of management applications.
The IT challenge stands to be enormous.
As a team led by Stepan Bogdan of Russia’s Tomsk Polytechnic University noted in a recent paper, “Today there is no ready-to-use framework applicable to make intelligent manufacturing systems for the oil and gas industry. “Network designers charged with bringing offshore production and performance data to the boardroom, along with remote management, will have to confront not merely familiar problems such as security, then, but develop an entirely new information architecture.”
That’s the aim of initiatives in Germany, known as Industry 4.0, and the United States, known as the Smart Manufacturing Leadership Coalition (SMLC). In both countries, the intent is to “overcome barriers to the development and deployment of Smart Manufacturing Systems through an implementation agenda for building a scaled, shared infrastructure called the Smart Manufacturing Platform,” as the SMLC describes it. Because offshore oil and gas is just now making the move toward comprehensive automation, it is likely to be among the first beneficiaries of the global research and movement toward smart factories, affecting operations, reporting those operations to management, and directing operations remotely.
Networking out breaches
Years ago, inventorying computer security at a best-unnamed manufacturing plant that was certain its network was secure, Tofino Security’s Chief Technology Officer Eric Byres found dozens of unguarded entrances to the system – some of them as obvious as an open USB port that any passerby could access.
When an enterprise spans the globe and its IT comprehends dozens of different systems with different architectures, in different locations with different office cultures, the hazard of data breaches increases correspondingly. Further, the remote location of offshore platforms makes cloud storage of data all but unavoidable, introducing still more potential vulnerabilities.
But traditional manufacturing MES/ERP were conceived for use within individual facilities, where all data and communications were local; consequently, many companies’ IT infrastructure and culture are unprepared to manage and defend the unending cascade of data that accompany doing business globally. In the oil and gas industry that problem is especially acute; as Bogdan notes: “Today, for most oil and gas companies, integration solutions are very expensive to support due to constant changes in business processes and engineering practices.”
Walmart, in the retail sector, mastered the trick of processing and responding to sales data from thousands of stores throughout the world two decades ago. But as Dieter Meuser, the CEO of MES-developer iTACS, points out, “Manufacturing IT is 20 years behind commercial IT in some respects.”
Catching-up is not merely a technology problem, but a cultural problem, too. What is about to change is that offshore platforms will soon cease to be distant reporting stations and become, instead, integral parts of the company’s central nervous system. With that, local security becomes a vital component of global security as well, inevitably requiring an increased emphasis on employee training and policing usage.
There is an additional danger. As security firm Symantec reported in its findings on June 30, an eastern European hacker network has been systematically attacking Western oil and gas facilities with software having a Stuxnet-like ability to seize control of industrial operating systems.
Russian hackers have been systematically targeting hundreds of Western oil and gas companies, as well as energy investment firms, Symantec said in its report. [ … ] The manner in which the hackers are targeting the companies also gives them the opportunity to seize control of industrial control systems from afar, in much the same way the United States and Israel were able to use the Stuxnet computer worm in 2009 to take control of an Iranian nuclear facility’s computer systems and destroy a fifth of the country’s uranium supply, the researchers said.
Universally, however, computer security experts are adamant: The biggest threats to security are not criminal hackers a half-dozen time zones away, but careless staff having legitimate authorization to access the network. As an offshore IT manager once told Offshore Engineer: “The number one issue we deal with isn’t keeping people out, but the people we have to let in.”
Though the displacement of manual labor by machinery is well-known to often put people out of work, it is also the case that improvements in productivity tend to lower costs and increase the demand for commodities, eventually expanding the overall demand for labor and driving payrolls upward. Just where the next generation of smart offshore work will fall in the age-old tussle between labor and management isn’t clear.
Undoubtedly, IT staffs will swell as companies develop and maintain global data networks, and design and fine-tune, over the course of years, the use of specific algorithms and software needed for parsing, interpreting, and allocating torrents of data.
Whether investment in MES, or more generally Smart Production, will conduce toward labor cost savings on offshore platforms isn’t clear. At least one recent study at a labor intensive machine shop found that “A definite reduction in labor hours can be realized by replacing current systems with MES, though as found, these benefits would merely reduce the time taken to perform tasks, rather than reduce the number of employees needed to perform them. Savings in labor overheads by replacing employees with MES start to become realizable only after exceeding 80% of the factory’s capacity utilization.”
Further, offshore oil and gas extraction have built-in boundaries that aren’t obviously susceptible to accommodating automation and automated management – unpredictable weather fluctuations, and variable product as recovery proceeds through rock layers formed by geologic processes millions of years ago. “The oil and gas business is forced to work within a techno-ecological system where there is a high level of uncertainty, and where controlling actions are usually indirect,” Bogdan said. However automation and distant, centralized management affect personnel numbers. What is certain is that even roughnecks are going to have to be better educated and attuned to modern technology than in times past.
The long-term personnel trends in offshore energy production then, both white-collar and blue-collar, will probably be shaped more by the global search for clean, non-carbon energy technologies than software, and the chief benefit of implementing MES/ERP and similar management tools will be to enable companies to remain competitive as margins shrink.
Toward an industrial epistemology
The architects of the Smart Factory paradigm are grappling with a question that has bedeviled philosophers for millennia: What does it mean to say that one “knows” something? Specifically, what does it mean to be sitting in a comfortable office somewhere on the Gulf Coast, staring at a screen full of neatly-arrayed numbers, and to say that one “knows” what is happening on dozens of production platforms hundreds, or even thousands, of miles away?
Numbers, however cleverly analyzed or intensively scrutinized, are neither experience nor institutional memory. From a study investigating how humans and machine-data interact:
“Emphasizing the duality of materiality in sensor-intensive work also questions the appropriateness of the idea that technology represents physical phenomena. Somewhat broadly speaking, both objectivist- and human-centric perspectives on knowing are grounded in the view that the knowing subject is ontologically separate from the object to be known. In such a view, representations mediate between the knowing subject and the object to be known. This ontological separation may explain why much of IS [information systems] research tends to focus on data/information representation on the one hand and human interpretation on the other as distinctly separate activities. The underlying assumption here is that representations are passive descriptions of reality. With basis in such a view, interpretation becomes the activity of decoding representations. The theory presented in this paper brings attention to how engineers constantly explore new ways of generating, calibrating, and visualizing sensor data by modifying hardware, software applications, and work activities. Such a sociomaterial conception of interpretation suggests that it is highly problematic to study representation and interpretation as distinctly different activities when studying the relationship between human activity and technology.”
|Siemens “future factory” in Amberg, Germany, is an integration of high-tech tools and high-tech workers that shines in stark contrast to the dark and dirty image of manufacturing in centuries past.|
It turns out that representation of data, and interpretation of data, are merely flip sides of the same coin: the way data is presented shapes the way data is interpreted and acted upon. The implication of that understanding, when fully realized, is that the Smart Factory and its instrumentation ought to aim not at being smooth-running but distinct from the humans who run them, but a smooth-running extension of the humans who run them, merging both into a single continuum of knowledge and reaction. That may not be so farfetched or science-fictionish as it sounds; according to an Industry Week report about the Siemens factory in Amberg, Germany:
“That future factory represents the absolute pinnacle of technological and manufacturing development, a perfect integration of high-tech tools and high-tech workers that shines in stark contrast to the dark and dirty image of manufacturing in centuries past.
“It’s a nice dream, that future. And it is a dream that, however fantastic, is much closer to reality than you think.
“The Siemens Electronic Works facility in Amberg, Germany, is a plant straight out of that dream.”
Similarly automated and interactive facilities have been built in India, and more are on the way.
For the oil and gas industry, Bogdan suggests the place to begin is with mining extant data as a basis for developing and implementing trustworthy heuristics, pointing toward well test data as an example:
“There is a problem of classifying well test results as valid or invalid. This classification depends on lots of factors such as random distortion, features of field exploitation, production intensification activities, production management style and others. So in practice geologists use complex expert estimation to make a decision: to use or not to use well test result data in later calculations. To support such decision making statistical models can be implemented. The goal is to create a model which can automatically estimate well test results and suggest validity of them using their history with predictive accuracy at least 75% for the most common test results.”
The Bogdan model achieved 82% agreement with the field geologists.
The first generation of offshore Smart Factory development will mine past performance data and strive to develop decision-making guidelines that emulate present practices, with constant feedback and refinement that ultimately matches or even exceeds human performance. The aim will be to eventually reach a fully integrated mesh of technologies and data-flow that runs smoothly and optimizes output with only minimal human participation.
Getting to NextGen
Transforming the existing stock of offshore production platforms to smart platforms integrally connected to the rest of the enterprise will need years of system upgrades, system integration, analysis, implementation – and a lot of money. Some older, marginally-producing platforms will doubtless be left behind.
For each platform brought forward, however, experienced software managers say that each installation should be treated as a separate, unique engineering project; it isn’t something you can do once, and then repeat with minor local tweaks. Every platform is unique, and so is its integration into the big picture.
Manufacturing Execution Systems: Optimal Design, Planning, and Deployment, a standard reference, emphasizes that failed implementations of large software projects are often a consequence of poor management. The place to begin, then, is by appointing a particular individual, and a deputy, who will manage the project, and “Both project managers must be released from other tasks in order to be able to dedicate their time completely to implementing the system.” The authors go on to identify another commonplace error:
“Two key points in the course of a project that are often given low priority (and thus too little time) are the specifications and employee qualifications. A guide for the time required for creating the specifications, especially for project-specific scope and interfaces, is the time scheduled for implementation. The time for specifications thus should be as long as the time allotted for implementation. Only this guarantees that there is sufficient time for including those involved in discussions and forming opinions. The release of production employees for training is difficult during ongoing production. Therefore, these dates and the related organizational regulations must be planned well ahead of time.”
The second key element in the planning is participation of the eventual users:
“Both wanting to and being able to work with the system are best encouraged through an open information policy and inclusion of employees in all phases of the project, where possible. An equipment operator can, for example, participate in creation of the specification in his or her area and can later support the system provider in the first test runs. System knowledge and acceptance are achieved automatically through this cooperation within the project. However, this cannot replace well-founded, systematic training. Therefore, training plans should be drawn up for all user groups.”
The all-important takeaway is that these projects are huge undertakings that need enterprise-wide buy-in and design collaboration. Anything less, and the transition to Industry 4.0 will end-up looking a lot like other big ideas that went bad: bloated, way over budget, and way behind schedule.
Bob Felton is a freelance writer based in Wake Forest, North Carolina.
Read more from OE Review's Networking and Connectivity special report.