June 2009

Tribal knowledge must go real time

By Bridget Fitzpatrick

Perhaps a little dramatic, but I have been known to tell young engineers to wonder as they enter a facility "what do I need to do today to escape certain death."

Certain death is certainly not likely to be imminent, but awareness is crucial. There is real danger in our professions. Accidents are rare, but things can and will go wrong. We all need to be able to respond rapidly and correctly to situations that we often have never seen.

Part of my concern stems from the reality that operators, engineers, managers, and even companies come and go, but the plants keep operating (or the ones that we work in at any rate).

In the U.S., the problem is perhaps more extreme than elsewhere. There is always change, but recently change is fast and furious. In addition, while we do not seem as concerned about it as we once were, we are also facing the large wave of retirements of the baby boomers. So it is not likely to get better any time soon.

In the good old days, the operators saw all the control loops all day as they walked the panel board. Engineers knew every loop by name and field location. Upsets were common. People gained muscle memory of how to respond to upsets from loss of instrument air and loss of power to all manner of mechanical and process troubles.

In the past, people commonly stayed at one company for their entire careers. They knew all the history. They knew what happened and why. They knew each other and how to work well through crises faced together. Transitions could be managed in a gradual and reasonable scale.

Looking at it simplistically, process control has gone through waves of evolution: local and then panel board controls to DCS and then APC/RTO/MES and the like. As we get better at process control, operators and engineers have less opportunity for experience in managing real upsets. Moreover, when the upsets happen, they are generally not simple; we have largely engineered away the one or two-dimensional upset.

There is undoubtedly a next wave of computing that will support smarter and more adaptive "learning" control systems. Control systems that readily support the complete information integration, real-time manufacturing, and truly high performance that competition demands are just past the horizon. However, they will not spring out of thin air. They will need some basis in reality, the correct constraints set from solid knowledge. This should be jazzy enough to attract a fresh crop of operators and engineers.

So here we sit, and we have some problems. For a variety of reasons, including the risk, the serial RIFs, and the notion that America is losing or has lost the manufacturing battle, the next generation seems to be less interested in the process industries. We need people. Most companies will say their people are their best asset. However, even the best among us expend relatively little energy in managing this resource. We manage career paths and product lines, but seldom manage the expertise of either in a way that is immediately useful.

I think we stand at a crossroads. Some will say the moment to turn the tide has passed. Nevertheless, I believe we can take back our future. We have to continue to learn to do more with less. To provide enough added value to financially take care of our best asset-people. We need to evolve in order to respond to real demands.

It used to be a business adage that you could not teach experience. However, we have no choice now. We have to develop technologies that give real-time advice, more commonly and with better fidelity than we do now. We have to expend the energy to build lessons learned into tribal knowledge in real time. This real-time nimbleness will result in better quality, better cost models.

We generally know what to do. How? Training simulators are becoming cost effective. APC models work to train operators and engineers. New software platforms are emerging. It is simply time to do the engineering work, take back the future of manufacturing, and reaffirm it as an American tradition.


Bridget Fitzpatrick (bridget.fitzpatrick@mustangeng.com) is a member of ISA and works on several standards committees. She is the HMI and Abnormal Condition Management Practice Lead for Mustang Engineering in Houston. She has degrees in chemical engineering and technology management.