of 6

PRINT DOCUMENT 6 - Alarm Management -Technical Brief

Published on July 2018 | Categories: Documents | Downloads: 14 | Comments: 0
148 views

Comments

Content

Technical Brief 

Establishing a Practical Alarm Management Philosophy By: Michael D. Johnson, P.E.

SCADA system alarms provide valuable warnings that help utilities know when important condition thresholds are exceeded and take preventive action before an emergency occurs. When managed correctly, alarms help ensure regulatory compliance and maintain conditions required to achieve the utility’s primary mission. Regrettably, poor management of alarm functionality often results in an overwhelming number of alarms, which causes operators to ignore early warnings or disable the signals altogether. In the course of automating and monitoring operations, control systems collect thousands, or sometimes millions, of data points about

many aspects of performance. These data points indicate status and condition of processes and equipment. During control system design, it is easy to identify many conditions that could trigger an alarm. However, more alarms do not always produce a well-managed operation. Like the authorities in the story of the boy who cried wolf, when inundated with to many alarms, people tend to tune out. When a legitimate alarm occurs it is often overlooked or lost, and an appropriate response is compromised. This article presents a proven process for  developing and implementing effective alarm management discipline and describes cases where alarm management was applied.

Technical Brief  One case describes a 165-mgd wastewater  treatment plant (WWTP) with more than 37,000 database points in its process control system. The other case study is 90-mgd water treatment and distribution system (WTDS) with over 5,000 database points in its SCADA system. Both agencies installed control systems that provided extensive control and alarm capabilities addressing nearly every aspect of the plant, including its processes and the performance and condition of its equipment. The systems displayed all alarms from plant processes and the computer system on operator alarm screens. However, this alarm capability generated far  more alarms than the system and operators could possibly manage. The sheer number of alarms – sometimes as frequent as one every 90 seconds – indicated a serious problem. No system-wide philosophy existed for managing alarms, nor were there any design guidelines or procedures for adding new alarms or for removing existing alarms. Nuisance alarms overwhelmed displays, burying critical alarms. Alarms occurred even when no operator action was needed. Many alarms activated even when nothing was wrong. At times, operators were not sure how to respond when alarms occurred. Some alarms were active for weeks or even months at a time. These factors could easily erupt into alarm "storms” that occurred when routine operations produced a large number of alarms, or when minor condition upsets generated significant numbers of alarms. Major operating upsets would result in an unmanageable number of  alarms.













Delay alarms are transient alarms that could be reduced by system functions, such as initial delay or cutout delay, that decrease the system's sensitivity to detecting transient alarms. Dead band alarms are "'chattering" alarms that could be reduced by changing an alarm's dead band, the small range around an alarm value within which an alarm's status does not change. Priority alarms are alarms that could be reassigned to a lower priority. Sensor-range alarms resulted from sensor ranges being set too close for  calibration tolerances. Limit alarms resulted from limits being set too close to normal operating parameters. Inhibit alarms resulted from normal operations (e.g. equipment designated as “out of service” or “standby”). These alarms could be inhibited by indicating to the control system that a normal operation had occurred.

This audit provided an initial description of the issues related to alarm management facing the operators. Redundant and inhibit alarms comprised 64% of  the nuisance alarms. Almost 10% were caused by alarm limits being set too close to acceptable operating limits or calibration tolerances.

Figure 1 – Causes of Nuisance Alarms

Auditing the Alarms  An initial audit of the alarm system at the WW TP found more than 700 active alarms, mostly in the nuisance category (Figure 1). Similarly, an audit of the WTDS found about 450 active alarms. An analysis at the WWTP defined seven categories of nuisance alarms and the percentage of  occurrence: •

Redundant alarms repeated or indicated the same root cause as other alarms.

Alarm Management Best Practices

Page 2

Technical Brief   Another 13% of the alarms were considered a lower priority than assigned.

Improving alarm management requires assessing, designing, implementing and optimizing alarms (Figure 2).

The process of improving the alarm system begins by developing a philosophy for managing alarms. The philosophy (Figure 3) considers normal operating limits, recognizing and avoiding hazardous situations, identifying deviations that could lead to financial loss, and improving understanding of complex processes. Fundamental alarm conditions are defined. For  example, a warning may be required when a process variable has passed a defined limit or is approaching an undesirable or unsafe value. The philosophy also considers how alarms are indicated (e.g. audible, visual, message to remote device, or a combination of these indicators). Most importantly, the philosophy establishes that operators must take action – and ultimately what that action should be.

The Assess step audits active alarms, analyzes root causes, and develops a philosophy for  managing alarms.

Figure 3 – Sample Philosophy for  Effective Alarm Management

The audit found active alarms on database points designed to record test data and monitor  equipment performance. Conducting tests and monitoring equipment performance are usually considered normal activities and typically do not require operator action. These points should be displayed only to the persons conducting the test or monitoring the equipment, rather than to all operators.

Following a Four-Step Process

The Design step documents guidelines for  determining what constitutes an alarm, and develops procedures for adding and removing alarms. The Implementation step applies design guidelines for alarm priorities, limits, delays and dead bands. This step also modifies operating procedures and provides training to support the alarm philosophy. The Optimize step uses analytical tools to investigate causes of specific alarms, develops graphic alarm presentation to aid in interpreting and responding to process upsets. Optimization includes regular reviews of alarms to continue minimizing nuisance alarms and address new alarm issues.

Figure 2 – Four-Step Process

Philosophy • •







Normal conditions should not produce alarms. Operators are shown only alarms they can act upon. Alarms alert operations to hazardous situations. Alarms identify deviations with potential financial loss. Alarms REQUIRE operator action.

Approach to Problem Alarms Problem alarms, those that don’t fit the philosophy, are resolved by: •

Removing out-of-service, redundant and useless alarms.



Reassignment of priorit y.



Application of filters, inhibit programs, delays.



Adjustment of dead bands and sensor ranges.



Activation of auto-acknowledge/reset features.

This shared understanding is embodied in the  Alarm Management Philosophy. Design guidelines are then developed to document when and how to add and remove alarms. The approach to managing alarms attempts to resolve the largest sources of problem alarms first. Often, these sources are resolved by reconfiguring alarm

Alarm Management Best Practices

Page 3

Technical Brief  processing parameters. The following actions are typically reconfiguration activities in the implementation step: •









Obsolete alarms are removed. Dead bands are increased to reduce alarm "chatter." Alarm delays are applied to eliminate alarms that simply note transient conditions. Priorities are changed to separate warning conditions from actual alarm conditions. Alarm-Inhibit functions are introduced to reduce alarm "flooding" resulting from normal operating conditions that generate alarms.

In both cases, a second audit was conducted one year later to measure the progress and results from implementing the Alarm Management Philosophy. At the WWTP, a comparison of the second alarm audit to the initial audit confirmed the approach was effective in reducing the largest sources of nuisance alarms (Figure 4). However, the total number of  nuisance alarms did not decline significantly. Ninety-six percent of the remaining nuisance alarm sources either were difficult to eliminate (requiring program modifications to inhibit alarms on a case-by-case basis) or hard to identify (requiring individual analysis of the alarm cause).

The second audit also demonstrated the need for more comprehensive analysis tools. At the WTDS, the comparison revealed a 40% reduction in the number of active alarms, due primarily to effectiveness of the reconfiguration activities.

Optimizing Analytical and Maintenance Procedures Optimizing alarm management often requires collecting and analyzing significant amounts of  "hard" data about alarm frequency and occurrences. Because previous analysis is based on reports of alarm occurrences, it often cannot provide clear insight into specific alarm deficiencies. A comprehensive analysis must capture and assess alarm activations, acknowledgements, actions taken, and system status events, all of which must be time-stamped to support statistical analysis. Because control systems can generate more than 1,000 alarm events per day, or an alarm every 90 seconds, a three-month history of  alarm events easily can include more than 200,000 events.

Figure 4 – Analysis of Alarm Sources after Applying Corrective Action

The value of the initial audit was in eliminating nuisance alarms which cleared the way to determine the primary source of the remaining alarms. In this case, the second audit was able to identify the root cause – a failed analyzer  which needed to be corrected. Ultimately this correction reduced the total number of alarms by an additional 40%.

Alarm Management Best Practices

Page 4

Technical Brief 

Figure 5 – Statistical Alarm Analysis

 At the WWTP, the analysis addressed the 10 alarms that occurred most frequently in a single week (Figure 5). This analysis revealed that a controller that monitored the temperature of  digester sludge activated 1,484 alarms – about once every seven minutes. Three hydrocarbon analyzers (all the same make, model, and application) generated 2,519 failure alarms. These four alarm sources contributed to more than 39% of the total alarm activations during the one-week assessment period. Since the three hydrocarbon analyzer failure alarms shared a similar cause, resolving these two alarm causes reduced the number of active alarms during the week by almost 40%. This sort of analysis helps those who manage the alarm system to quickly identify, prioritize and resolve specific problem alarms.

Conclusion Modern process control systems provide tremendous capacity to display alarms related to an enormous number of plant processes and conditions. For these functions to be meaningful – and ultimately improve the utility performance – this capability must be planned, implemented and managed. By applying the proven engineering approach of  assessing, designing, implementing, and optimizing, we can ensure effective alarm management, reduce active alarms, eliminate nuisance alarms, reduce unacknowledged alarms to zero, and significantly improve early detection and prevention of discharge violations.

When the philosophy has been applied consistently, icons can used in new alarm screens to further optimize alarm presentation and operator response.

Alarm Management Best Practices

Page 5

Technical Brief  Additional Information For additional information contact: Michael Johnson, P.E. Regional Principal Engineer  Westin Engineering, Inc. 3100 Zinfandel Drive, Suite 300 Rancho Cordova, CA 95670

(916) 208-0543 [email protected] www.we-inc.com

 At Westin, we're passionate about improving utility performance. Through practical knowledge earned in the water and wastewater industry and our  collaborative approach, Westin helps you implement dependable enterprise solutions to retain knowledge, optimize processes, meet compliance requirements, serve customers and improve asset life.

Alarm Management Best Practices

Page 6

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close