To advertise on this website,
please write to
Bill Hollifield, PAS Principal Alarm Management and HMI Consultant and Randy Cole, PAS Director Technology Applications – Power, discuss the need for effective alarm systems in infrastructure facilities
The power sector in the Middle East is experiencing record-setting growth in both demand and capacity. Significant population growth, industrialisation and construction are the main factors influencing this trend. Some analysts expect capacity growth to be approximately seven to eight per cent throughout the region, higher in some countries, and lasting for many years.
Private investment strategies in some countries are attracting significant development of new capacity. These projects are accomplished quickly and efficiently by large engineering and construction companies but unfortunately, these companies have little to no actual operating experience and the plants they produce are often copied from previous designs.
In practice, those designs have been shown to be seriously deficient in a very important area – the alarm system used by plant operators.
Contrary to its purpose, a poorly configured and performing alarm system in a power plant can actually hinder the operator’s ability to effectively manage abnormal situations. Increasingly, we have found that alarm system failures are a contributing factor to power generation incidents and accidents.
These are common problems throughout the power industry and have resulted in unplanned outages and decreased profitability. An effective alarm system is a key element for reliable power generation operations.
By exploring the origins of the problem, its nature, and applying a proven seven-step methodology, we will illustrate how significant improvements to both existing alarm systems and new designs can be made.
The widespread use of modern Distributed Control Systems (DCSs) is the catalyst for today’s alarm system problems. Effective alarm system design best practices were unknown when most of these systems were initially installed, resulting in inconsistent application and a near-exponential growth in the number of configured alarms.
The creation of alarms in modern control systems is performed through software configuration and has no incremental cost. Prior to this, adding an alarm involved the installation of hardware and the costs associated with that, which ensured that every alarm was justified and meaningful.
It is not unusual to see a single operator console with 3,500+ configured alarms, which can result in hundreds to thousands of annunciated alarm occurrences per day. This overwhelms the operator, making it impossible to acknowledge and respond to each one,
and making it easy to miss the important notifications.
In the power industry, the alarm system configuration is often improperly carried out by system integrators, without much involvement of the end user of the system. Poor initial alarm configurations can result from the reuse of engineering from prior designs or using rules of thumb. As such, the alarm system will not prioritise alarms correctly and they may not be consistently classified, causing trivial alarms and important situations to appear equal to the operator.
Proper routing of alarms can also improve other aspects of plant reliability. For example, routing the relevant low-priority alarms to a predictive maintenance system can help detect trends and pre-empt future problems.
Another common problem is that many companies do not have proper alarm system management-of-change (MOC) policies. Some allow operators to change alarm settings at their individual preference, or to suppress their annunciation, without documentation or proper consideration of engineering design.
Such practices cause an alarm system to change almost randomly over time and become even more inconsistent – especially when operations staff move on. Some of the most prevalent alarm system problems include:
• High continuous alarm rates – Alarm rates are often far above the ability of an operator to handle. Thousands of alarms must be ignored each week in such a system, with no guarantee that the right ones are always acted upon.
• Alarm floods – An operator may experience hundreds of alarms within a few minutes of a minor upset, which may mask critical alarms and prevent timely corrective action.
• Improperly suppressed alarms – Without records or notifications, improperly suppressed alarms can indefinitely eliminate the annunciation of other important alarms.
• Chattering and nuisance alarms – Such inappropriate alarm behaviours contribute to operator fatigue and make detection of valid alarms more difficult.
• Stale or long-standing alarms – These clutter the alarm system, also making detection of valid alarms difficult.
Article continues on next page ...