In this, and future columns, I'll be making some observations on current trends in process control. While 90% of global production is still controlled by analog instrumentation, almost all the controls installed as a part of a new plant or plant expansion are digital control systems (DCS) connected by digital networks.
In the 1970s I designed a control system for IBM's headquarters in such a way that the building would heat itself while eliminating the chimney effect. Although the building was 43 stories tall, the control room (naturally) was in the third basement. During startup, communication became extremely difficult. How difficult? Just imagine that you have a "hang up" in a herding-optimization scheme, because there's a valve sticking on the top floor. To fix it, you need to know the response of the loop, but that's displayed 46 floors down.
Nowadays, in this era of digital buses, one can plug in a laptop or use a wireless hand tool to instantly establish access to all the data, displays and intelligence that resides anywhere on the DCS network. This capability, in combination with the self-tuning, self-diagnosing and optimizing features of modern process control, makes both startup activity and operational routines much easier and more efficient.
Similarly, DCS's offer process modeling and simulation, something that can improve operator training a great deal. An accurate simulation model allows operators to train under "live" conditions without exposing the plant to the consequences of their mistakes .
However, these positive aspects are only part of the picture and, on the other side of the coin, there are some serious issues you might have to contend with. A big problem many are encountering these days is the tendency for unscrupulous DCS vendors to "sell their violins without strings." They do this by neglecting to identify all the software needed to operate the system, treat software as an extra, and exclude the cost of preparing the unique control algorithms, faceplates and graphic displays in the basic bid. All of which can lead to some serious project management problems.
It's been my experience that if the plant does not hire an engineering firm or system integrator to "string" the violin (DCS), and relies solely on plant personnel to do it, the control system will be woefully out of tune. Be aware though that to fill in the software preparation gap, the associated costs may exceed the entire system's total hardware costs.
The bid language can also get pretty murky. In some bids, one might read that the stated cost is for "hardware with software license." To some, this implies that the operating software for the DCS package is included.
In many cases, unfortunately, it is not; only the license to use the software is included. Similarly, when you read that an analyzer, optimization or simulation package needs "layering," or is in the "eighth layer," one might think that the bid contains eight device layers. What this often means is that the cost of integrating these packages into the overall control system is an extra.
So, on the one hand, the age of the plant-wide digital networks and associated advanced controls have opened the door for great optimization opportunities. On the other, though, it requires a concerted effort before all the pieces of the puzzle fit together, and the "violins" can orchestrate automation harmony.
In the coming months, look for my columns to focus on connectivity and system integration, advanced control algorithms and optimization strategies, and finally, some advice on managing overall DCS costs.
BÂ©la LiptÂ¡k, PE, process control consultant, is also editor of the Instrument Engineers' Handbook and is eeking new co-authors for the forthcoming new edition of that mult-volume work. He can be reached at firstname.lastname@example.org.