Effective Maintenance Management. V. Narayan
necessary to keep us alert, active, and expectant. We call this facilitative stress. Too high a stress level can be as a result of physical or psychological pressures. This may result in tiredness and lack of concentration.Too low a stress can be due to the work being repetitive, intellectually undemanding, or otherwise boring. During World War II, the British Royal Navy noted that submarine lookouts became ineffective after about 30 minutes, as they could not remain alert. The lookouts knew that their own lives depended on their vigilance, so motivation was not an issue.
Swain and Guttmann4 give the following examples of psychological stress:
•Suddenness of onset
•Duration of stress
•Task speed
•Task load
•High jeopardy risk
•Threats of failure, loss of job
•Monotonous, degrading, or meaningless work
•Conflicts of motives about job performance
•Reinforcement absent or negative
•Sensory deprivation
•Distractions such as noise, glare, flicker, color, or movement
•Inconsistent cueing
Each person is slightly different and thrives under different levels of stress. However, a number of the stress factors affect many people in similar ways.
In order to reduce human failures, we have to address the factors contributing to stress. By doing so, we can produce the right environment for each person. In most cases, we will not be able to influence stress caused by domestic matters, so we will focus on those at work. Job enrichment deals with the elimination of boredom and unacceptably low stress levels. We can attribute the remaining problems to high stress at work.
Control room operators perform critical functions. During plant upsets, startups, and shutdowns, their skills are in demand. We use alarms to catch their attention when things go wrong. Designers of control rooms have to take care to minimize the number of alarms they install. If too many alarms come on too quickly during a plant upset, operators can lose concentration and react incorrectly, thereby worsening the situation. In an article entitled ‘How Alarming!,’ Bransby and Jenkinson5 report the results of a survey. They studied 96 control room operators in 13 different plants in the U.K. Their findings, listed below, indicate that we have to devote more attention to this issue at the design stage.
•In an average shift, during steady operations, operators receive an alarm about every two minutes;
•Many of these alarms repeats of ones that occurred in the previous five minutes;
•Operators stated that many of them were of little value to them, and that eliminating about 50% would have little or no effect;
•Following a plant disturbance, they estimated that there were about 90 alarms in the first minute and seventy in the next ten minutes;
•About half the operators said that they felt forced to accept alarms during plant upsets, without reading or understanding them;
•During the survey, they observed one such plant upset. The operator did not make a full check of the alarms for about half an hour. This behavior was consistent with that reported by the others in the survey.
Because the purpose of the alarm is to alert the operator, these results indicate that the designers have failed in their objectives. The authors state that improvements are possible, and that a variety of tools are available. Some of the simpler ones include tuning up limit values and dead bands, and adjusting priorities. The use of logic to suppress some non essential posttrip alarms is also possible. As an example, they state that a review of the alarms resulted in a 30% reduction in the number of alarms. A structured and logical process is available to manage instrumented protective systems, which can help designers optimize the number of alarms and trips. We will discuss this process further in Chapter 10.
One of the causes of human failures is tiredness, and this is often due to sleep deprivation. The human body operates with the help of a biological clock. Shift work can disturb normal (or circadian) sleep cycles. As a result, the reaction to stimuli can be slow. This can affect the ability of the operator to respond to a rapidly developing scenario. Night shift workers are more susceptible to this problem than the rest, because of the disturbance to their circadian rhythm. Although there is no direct cause and effect relationship established, we note that some of the worst industrial disasters including Piper Alpha, Bhopal, Chernobyl, Three Mile Island, and Exxon Valdez occurred in the silent hours. This does not automatically mean that it is unsafe to work at night. Night shift workers have completed many millions of hours of work without any incidents. It is the combination of circumstances that matter, so one must view this in context. Because we cannot eliminate night shift work, especially in continuous process plants, we have to try to understand the risks, so that we can take suitable steps to minimize them.
A factor affecting sleep cycles is the way we arrange shift patterns. Lardner and Miles6 have explained why some shift patterns are superior to others from an ergonomic point of view. They propose a nine day cycle, with 2 days each in the morning, afternoon, and night shifts, with a 3 day ‘weekend’ following the night shift. The weekend may turn out to be in the middle of the week. They argue that this pattern is superior to the alternative 28 day cycle, which is quite common. The 28 day cycle consists of 7 night shifts and 7 evening shifts, followed by a 2 day weekend after each block. This sequence followed by 7 morning shifts and a 3 day weekend.
Human errors occur due to a number of reasons, and lack of knowledge and experience are not necessarily the most common. Motivation and morale are often key issues to manage. Pride in work, a sense of being wanted, and being treated fairly are all important considerations. We all want user-friendly software; similarly, workers appreciate managers who are people friendly.When this is so, we are likely to experience lower absenteeism or sickness, better participation in team effort and suggestion schemes, lower accident rates, and higher productivity.
What makes human beings distinctly different from machines is their ability to think, often in a very creative manner. Feelings and emotions change the way a person responds to identical stimuli over time, and makes it hard to predict behavior. We have provided a brief introduction to the subject in this chapter and readers can refer to Lorenzo’s excellent guide for a more detailed discussion. A check-list of potential causes of human errors is available in Appendix 4-1.
We began this chapter by defining failure in relation to the required performance standards. Failures can be critical (causing total loss of function), degraded (where the loss is partial), or incipient (where progressive deterioration has commenced, but will take some time before there is loss of function). We note the significance of the operating context, and how this explains why identical items of equipment perform differently. We saw how failures themselves provided a means of control of the process.
Our next topic was the role of maintenance in achieving the desired equipment performance. We discussed incipiency, and its use in conditionbased maintenance, using some common examples to illustrate the concepts. Thereafter, we discussed age-related failures.
Finally, we looked at human errors, perhaps the most complex issue relating to failures. We noted that there is an optimum level of stress required to keep human errors as low as possible. The work done by experts on sleep cycles shows us how they can affect the body’s natural rhythm. The experts state that some shift patterns are superior to others when planning 24-hour coverage for continuous process plants.
Feelings and emotions play a major role in affecting the way people react to situations. Therefore, managers have to focus on motivation and morale, which are key issues in minimizing human failures.