Knowledge mining drubs the competition

Issue 5 and Volume 101.

Knowledge mining drubs the competition

By Timothy B. DeMoss,Associate Editor

Several years ago, I remember thinking there was no way technological advances could keep up their current pace. Graphs depicting these advances swung upward at the end like a stomach-churning ride at Six Flags. Yet today, that trend remains unchanged. We are now overwhelmed at times by the flurry of technologies reported in the news and pushed by marketers. It`s amazing that compared to what is on the store shelf today, my two-year old computer, despite my feeble upgrades, is more suitable as a planter than a processor. Nonetheless, with the proper software and hardware attachments on each end, I could use that `old` computer to make a free long-distance call to Scotland and ask Ian Wilmut to show me a live picture of his cloned sheep, Dolly. Even medical advances have been progressing at an astonishing rate. The world has come far since Alexander Graham Bell spoke into the first phone and said, “Come here Watson, bring me some aspirin.” Or something like that.

As engineers, it is our job to sift through all the new technology and apply it in unique, innovative ways to make the world function better. To power plant engineers, new technology often means more efficient turbines or better pollution controls. Plant controls and automation certainly belong in the communications and information technology category. But there is no reason we can`t use “common” technologies, like the one you could use to call Scotland over the Internet, to boost our efficiencies as engineers.

Team player

Make fun of workplace teams if you like; but when people choose to work as a team, the results can be excellent. Not everyone works in Dilbert`s office. Of course, a team`s potential relies solely on the abilities and collective knowledge of its members. This is significant today because for the first time in history, the technology is widely available to enable collaboration by groups of people in different geographic locations. For years we`ve been able to collaborate over the phone. However, the phone has its limitations. It does not lend itself to long-term collaboration on a difficult problem by dozens or hundreds of people. It`s hardly the next best thing to being there. Today`s technologies, on the other hand, solve not only the large-team problem, but the “being there” problem as well. Although the following examples reference Internet technology, the same efforts could be applied within a utility`s Intranet.

Large-team technology

One popular form of communication on the Internet is “chat,” by which people can have real-time conversations by typing messages on their computers. These messages are sent to a chat server, which then makes them available to read by anyone “tuned in” to that particular chat channel. For small groups of people, chat can be a decent, inexpensive collaboration tool. However, in the corporate world, where long-distance conference calls can be made for pennies per minute, this format is not a viable alternative. In this scenario, typing your comments becomes an obstruction. E-mail is the most used application and most used form of communication on the Internet. And while it is useful for one-to-one communication, it also provides a cheap, quick way to bring together dozens or hundreds of people in a many-to-many conversation. Typing messages in real time is impossible with this format, but it does seem to lead to more reflection before commenting by the participants.

Using an e-mail listserver, which routes each e-mail it receives to every participant, many organizations find creative solutions to problems by tapping the brain resources of everyone involved, while fostering a sense of community or team within the organization. Many companies use e-mail only for mundane communications, overlooking the possibility of forming even temporary e-mail listserver groups to tackle complex problems which otherwise would be tackled by only a few people in that most efficient of settings (ahem), the conference room meeting. Certainly the e-mail approach requires good leadership to put the pieces together in a formalized solution, but that is no different than in traditional approaches.

Another approach similar to using e-mail is to use a bulletin-board or newsgroup-type discussion list. The main difference between this type of collaboration tool and e-mail is the users` ability to read the messages grouped by subject, or threaded. This approach is becoming popular with many companies. At Texas Utilities Electric Co. (TU), employees have been using on-line discussion forums for more than a year to discuss a number of topics related to the utility. There have been Q&A about employee benefits, announcements for new Intranet services offered by various departments, and of course, requests for help on Net questions or problems. According to Chuck Preecs, aTU product development engineer who maintains the misc.industry.utilities.giflectric Web site (www.webfeats.com/preecs/miue) when he is not hard at work at TU, “the system has been used some, although not a lot, primarily due to the lack of Intranet access by a large number of employees and the shift it represents in the way people are accustomed to communicating within the company (usually by phone, person to person, etc.).” “However,” Preecs said, “use is growing steadily over time.”

Although e-mail and message boards seem primitive compared to live chat boards or videoconferencing, they provide a built-in filter for the cacophony of “voices” with which one must contend in a real-time collaboration.

Creating new knowledge

E-mail, newsgroup and chat may not be hot news to many of you. Disadvantages to real-time collaboration aside, however, the day has arrived when technology has made it possible to collaborate with colleagues in a different building or a different country in real time, without the visual and audio restrictions placed on us by text-based-only communication. The newest generation of communications tools even provide the ability for more than one person to manipulate computer files simultaneously. There are a handful (soon to be dozens) of software and hardware packages available that allow multiple users to view and make changes to CAD drawings, word processing documents, spreadsheets or any other document you can call up on your monitor. As more of these programs are developed in cross-platform languages like Java, you will no longer even have to worry about massive upgrades to your existing computer infrastructure.

Combining this new capability and other electronic forms of communication and collaboration, with the network infrastructure now in place to utilize it, brings group collaboration into a new age. Many labels have been applied to this technology, but the one used most often is “groupware.” Some of the bigger companies involved in bringing groupware to the masses should sound familiar (Microsoft, Corel, Lotus, IBM, Digital Equipment Corp., Netscape). Much has been written about groupware (check out www.crew.umich.gifdu:80/~brinck/cscw.html) over the last few years. But there seems to be a common ground in assessing its worth. Britton Manasco, The Learning Enterprise editor, said it this way: “What is most striking about the technology is not its ability to disseminate existing information, but rather, its potential to facilitate the creation of new knowledge in a collaborative context and help companies better manage their intellectual assets.” It is this creation of new knowledge based on shared knowledge and experience that will be essential to power producers as we head through the valley of the shadow of deregulation. Relying on old guard information management and communication will not be enough when the competition heats up.

Learning by example

Industry-wide collaboration can also help foster and nurture new technologies that will be useful to all. The medical field has latched onto electronic collaboration with gusto because the complexity of the problems that arise there and the projects undertaken (the Human Genome Project, for example) are perfect matches for groupware technology. If the medical community, despite the ethical questions associated with publicly sharing patient information, can make strides in curing diseases or cracking the code to our existence, the electric industry should be able to tackle problems like environmental regulation and sustainability in a similar manner while still realizing healthy profits. Some companies are leading the way to collaborative power engineering (see “Solving the burden of power plant documentation,” page 26), but it is up to the plant managers and engineers to use these kinds of tools effectively.

What`s hot

What we`ve failed to address thus far is where the different communications and information technologies fit in the overall corporate picture. Because these technologies are so new, few companies have developed a cogent strategy to implement them. However, those companies which have begun to paint the big picture are developing a communications plan similar to that depicted in the figure on page 24. It is useful to think of the concentric rings in the figure as layers of security which become tighter as they approach the center. For power plant communications, most of the technologies and ideas covered previously would be relegated to the utility Intranet, available only to the local employees. Companies normally reserve the most secure level for company officers or system administrators, although these two groups would rarely have access to each other`s domain.

The figure includes the wide-area network (WAN) as a separate level. However, that level may only exist when geographically distant employees access the system through public means such as an Internet access provider. WANs with a dedicated, secure-line infrastructure would operate just like an Intranet. The extranet provides a level of security which allows entry to vendors or other groups who need access to more than just publicly available information. For example, a company might use an extranet to allow electronic document transfer of invoices or to allow established customers to access information regarding account balances or products in stock. Online banking and bill payment are services which use this concept. Finally, the Internet level would be open to anyone.

Although there are distinguishable levels of security, each level will rarely be distinct. Each company`s needs will dictate exactly which groups of people have access to the various levels. Again, not all of these security levels relate to power plant operations, but at the utility-wide level, such an arrangement forms the basis for a well integrated communications and information system. At POWER-GEN `96 in Orlando, Fla., many utilities discussed their intentions to build enterprise-wide information systems that coincide with the inner layers of the figure. Adding the extranet and Internet simply completes the concept.

Future control

As technologies continue to advance, allowing us to track more and more information, a problem arises around how to handle this information increase. In power plants, one`s ability to monitor every conceivable piece of equipment brings more information to the table than most operators can handle without the help of computerized controls. These control aids do a respectable job of corralling information overload, but as efficiency`s role becomes even more prominent with deregulation, utilities will undoubtedly look again to new technologies that can produce a sweeter bottom line.

New mechanical technologies and systems will certainly play a part, but utilities should begin to look to advanced information control technologies as a next possible step toward squeezing more points out of the efficiency equation. In this case, efficiency can refer to shift scheduling just as easily as turbine performance. If employees are a company`s biggest asset, then the information in their brains is what makes them valuable. If workers can be freed from mundane tasks to concentrate on more complex issues, then the overall efficiency (output/input) of the operation should increase.

Q: How will the utility of the future leverage this asset to its advantage?

A: Artificial intelligence (AI).

Artificial intelligence

The concept of using AI to run a more efficient utility is not as far-fetched as it seems. Although it is not available today for the electric power industry, in financial circles what was once scoffed at by “legitimate” traders is now making some people very rich. The similarities between running a power plant and trading financial properties in the world market may at first glance seem marginal. However, the complex interaction and correlations among and between the subtle changes in a financial market are no more complex than the relationships among those things involved in running a power plant.

The effect that changing the fuel mix for a gas turbine has on the NOx emissions from the stack is just as suited to an AI application as the effect the European bond market has on the stock price for General Motors. An article in the December 1996 issue of Wired magazine highlighted a British financial firm, Pareto Partners Ltd., which “uploaded” the knowledge of one of its star traders into an AI system, developed at Hughes Electronics Corp. in cooperation with Pareto. On some level, it is like brain cloning. Unlike some of its AI predecessors, Hughes` Modular Knowledge Acquisition Toolkit has been highly successful for Pareto.

While Pareto`s star trader possessed the knowledge of how individual subtleties in the market affect other market details, her inability to track more than a handful of these variables at one time had prevented her from making more than just educated guesses about what actions to take as a result of examining the data. AI made it possible to evaluate huge numbers of market variables using the trader`s market knowledge to make decisions based on available data. Likewise, in a power plant with hundreds of linked variables, the number of relationship combinations approaches grotesque proportions. Factor in job stress, fatigue, emotions and other human distractions and the prospect of using AI to make operating decisions begins to look inviting. Consider that a power plant`s variable correlations are more exact than those deriving from economic theory, and power plant AI begins to look like a possibility.

Power plant AI may be some time away, but the community building and collaboration mentioned previously are important steps to making it a reality. Using modern computer technology to foster communication and manage information, followed by using the computer to help sort out the information available to us is an exciting possibility for the industry, and one we should pursue. As Martha Stewart might say, “It`s a good thing.” z

Did you find this article interesting?

Yes: Circle 300 No: Circle 301