Manage process control for maximum success

01 February 2006

Proper planning, organising, monitoring, and correcting are traditional values that help shape success. In today’s short-term, competitive world of the process industries, they should not be overlooked.

Process control work is often driven by spontaneous, temporary demand, to fix a problem here or there. That is in the nature of this service. Even the development of larger, Advanced Process Control (APC) applications, an activity that takes more than just a few hours, is often triggered by an actual need and not by an application implementation plan.

Responding to actual plant needs is and will be always a major portion of the process control activities, yet sound organising, planning and follow up can have its merits, especially for longer term work. The question we want to deal with today is what measures could help to deliver better solutions and to save time as well.

Lets us start by looking at the key success factors for process control and especially for APC and then see how we can contribute to them. In the English language they can be described as the four T’s, namely

....Training;
....Technology;
....Tools; and
....Tactics.

Training—The importance of good education was never disputed (although here and there maybe a bit overlooked). It became really visible by the often cited Solomon & Associates study where the intensity of the use of APC and the education level of the staff were found to be the two statistically most significant
contributors to the profitability of the investigated oil refineries. It is not too difficult to imagine that this is very likely true for the entire process industry. Of course, training has to be planned in line with the applications development and thus should be part of the overall master plan—which we are going to look at later.

Technology—This is a most obvious enabler for improved control and thus operations performance. The key
question here is, which are the most suited techniques and approaches for the specific situation? The point to be made is that performance is only one aspect in this respect. The second one is just as important, namely the utilisation of the application, in other words its ‘Service Factor.’

With respect to the Service Factor, any candidate technology should be a) preferentially usable right in the DCS; b) as simple and robust as possible (in the sense of the famous Einstein quote: ‘All models should be as simple as possible— but no simpler!’); and c) easy to understand to ensure fast (or even better: immediate) operator acceptance.

Tools—Manpower and time are in short supply and tools are playing an increasing importance in the quest to deliver better solutions in shorter time. Specialised software packages for visualisation and analysis of operations data, process parameters estimation, PID tuning calculation, decision on the structure of the control scheme (e.g. the use of a cascade or a feedforward) and statistics for trouble shooting are the
among the main candidates. Once the future control applications are defined, then the appropriate toolset can be selected. The emphasis should be not only on the power but on the ease and consistency of use.

Tactics—Under this factor there are all the strategies and plans that could and should be in place to ensure that work is carried out in the most effective and efficient way. It is a wide field, so let us focus on those items that are important yet, in many cases, deserve improvement. The organisation While there are many organisational issues, two are most frequently encountered.

First, staff who are charged with tasks that take more than just a couple days, like the development of an APC application, should be totally detached from daily business. They should be allowed to fully concentrate on developing the applications. Often the need for better control is identified and a suitable control strategy is developed in a very short amount of time, but when it comes to implementation, plant tests, controller
tuning, and so forth, progress can be extremely slow if the involved persons are already dragged into other tasks. The original ‘hot’ project will simmer on the ‘back burner.’ In short: Mingling APC and daily business does not work.

Secondly, application design and implementation should be done by the same people. Although this sounds selfevident it is surprisingly common in many organisations that one group, who has no DCS skills, develops the application, while another group, often IT folks with no control background, is responsible for the implementation. This commonly leads to problems arising from different priorities, lack of understanding and communication. Carrying out all the tasks by one person or group of persons implies
that they must have the necessary DCS skills. This may require extra training, but in contrast to the payout, this is a relatively small investment.

DCS vendor selection
The decision regarding the vendor and the system is by no means trivial, especially when different operations (for example, continuous vs. batch) and requirements (pure automation vs. APC) need to be covered. Yet it pays to go through a thorough investigation of the company’s needs with the goal to standardise on one or two vendors or systems.

Of course, this is even more difficult where different organisations are involved. At sites where several business units are represented it is quite typical that several different systems are installed. The maximum number I have seen on a mid-size site is seven!. The disadvantage of such a configuration becomes more evident in serious situations: With staffing in every unit already at or below the critical level often help from other units is needed—but almost impossible to deliver. Furthermore, experience exchange, transfer of operator or control staff it is most cumbersome and costly and the position vis-a-vis the DCS vendor for
every unit quite weak.

Knowledge management (KM)
Building up, distributing, sharing, and using knowledge can have tremendous merits—as we have heard at all these KM conferences. Yet, very few human networks in the field of process control have been set up. As a consequence it is common experience that the outside consultant often knows more what the sister plant next door is doing than the company staff itself. There is nothing to be said against ad-hoc experience
exchange, but organised regular contacts, meetings and special focused joint activities simply will deliver more.

Performance monitoring and reporting
A crucial part of any management activity is of course the control of the action to be carried out. Since we are dealing with dynamic processes any meaningful monitoring must be on a dynamic basis, and not just based on daily, weekly, or monthly averages. Yet it is surprising how little sound knowledge exists about the
current status and performance of controls and operation in terms of variance, process capability, analyser
availability etc. Controllers and applications cost money and consequently their state and degree of utilisation must be closely observed in order to ensure that this investment is fully exploited.

Master plan
The overall guiding instrument for all these activities should be a ‘master plan.’ The master plan is based on a study of the plant with the objective of identifying all potential future control schemes and all necessary changes to the existing ones. The result is typically compiled into an Application Design Basis document. This document describes every application, its objective, inputs, outputs, any required equipment (new or different measurements analysers etc.), the technology used, and, of course wherever possible, the associated incentives in a brief, concise way.

On this basis an implementation plan can be developed as well as the needed equipment and new technology acquired, control staff trained in a timely synchronised fashion, and so on. The advantages are obvious, yet such long-term planning has become a bit out of fashion. Lack of time and manpower is often cited as the reason, but it should be noted that, against common belief, such a study does not necessary take enormous effort. In most cases it is a matter of a few man-weeks.

Studies (again, in oil refineries: this is the best data source we have) have shown that the average, occasional users of APC gain some extra 1$/ tonne profit from these activities. But they have also shown that the top users, the leading ‘beneficiaries’ are harvesting over 3$ / tonne, which for a mid-sized plant
amounts to more than $10m per year. These credits do not only come from the use of more advanced approaches and technologies but also from proper consideration and implementation of all the relevant organisational aspects. Proper planning, organising, monitoring, and correcting is a key element for
success—just as in any discipline.

Hans Eder, actgmbh@compuserve.com


Related Articles...

Additional Information...

Print this page | E-mail this page