Industrial Accidents 2.1 Accidents
2.5 Performance Indicators
Figure 2.45 RCA of the Bouncefield explosion developed by company Governors BV (NL).
Source: Adapted from [20]. Reproduced with permission.
of the industrial accidents are the natural and predictable consequence of a potential situation where some performance indicators have been misunderstood or undervalued.
A distinction between “lagging” indicators (related to actual loss events) and “leading” indicators (related to the precursor events) is done. The Lagging Process Safety Performance Indicator measures the number of process incidents. The term “lagging” stands for
“happened”; therefore, the Lagging Indicators are related to events which have already happened (incidents), with their consequences and the suggested corrective actions. Process incidental events are
classified and communicated according to the internal procedures of the company. On the other hand, the Leading Process Safety
Performance Indicator measures the effectiveness of the process safety management activities. The term “leading” stands for “precursor”; so, it refers to activities, factors, and parameters related to events that did not happen. Therefore, the Leading Indicators allow the identification of preventive measures for incidents.
Some lagging indicators are:
Major incident counts;
monetary losses;
injury/illness rates; and process safety incident rates.
On the other hand, the leading indicators include:
Near misses;
abnormal situations (like overpressure relief events, safety alarm or shutdown system actuation, flammable gas detector trips);
unsafe acts and conditions; and other PSM element metrics.
Safety performance measurement and monitoring require the
definition of some safety performance metrics, like the incident based metrics and the related statistical measures cited in [23]. Examples of performance indicators are:
Number of relevant safety recommendations, still open on the site.
It is the number of recommendations identified during the PHA, safety audits, incident investigation audits, near miss investigation audit, or similar activities. These recommendations have been evaluated as a priority, because they are related to high risks, and require an immediate action (within 1 year). The expectation on the evolution of this indicator is a decrease of the number of open
recommendation in time, until zero. The number and velocity of resolution of the indicator must be constantly monitored. On a regular basis, the number of still open recommendations has to be notified;
number of scheduled inspections and maintenance activities identified on safety critical equipment, with a delay of 60 days to be realised and without formal approval by the Direction. The indicator should distinguish the different equipment typologies;
reaching of the Safe Operating Limit. The indicator includes:
emergency automatic depressurizations, activations of safety valves and rupture discs, activation of safety interlocks, activation of
cooling system (for exothermic reactions);
training of plant operators. The indicator is expressed in hours/month for a single operator;
Percent Evaluation Sheet. The indicator expresses, in percent, the execution of safety tours respect to the goal number;
percentage of involvement in emergency training. The indicator expresses the annual frequency calculated as the number of emergency training executed by a single operator respect to the total expected number;
number of spurious activation of SIF;
number of bypasses in use over a process safeguard;
number of bypasses in use for more than 30 days without implementing change (MOC);
Permit to work. The percent of permits to work correctly closed;
number of the “Pre Start up Safety Review” not completed before
the startup;
number of implemented management changes without adherence with internal procedures; and
number of simulated emergency tests carried out over one year.
They are a measure about how the Health, Safety and Environment company's policy, which is established from the corporate
management and affects up to the line level, is effective in preventing and protecting from an incident.
The development of a performance indicator is discussed in [47], which identifies ten steps:
Identify and record the business outcome;
identify the process flow and record the process outcome;
identify and record the process purpose;
identify and record the most important outputs of the process;
identify the critical stages of the process and the dimensions of process performance; develop and record the measurements for each dimension;
develop and record goals for each measure. Goals must be specific, measurable, achievable, relevant and time sensitive; and
define and record the levels of success, indicating if the results have been achieved.
The incident analysis reveals a reoccurrence of some organisational factors playing an important role in triggering/causing/developing the incident itself. This identification process requires the previous
definition of a classification meter for those factors. This approach may be a bit difficult, since it is not easy to establish, a priori, which are the organisational phenomena that may result, in the end, to an incident. Indeed, the linking between those factors and the undesired outcomes is not direct: they are often interlinked, and a latent period of time is very often observed between the occurrence of the factors and the one of the incident. This means that it is not possible to
establish, a priori, a chronological order and therefore it is impossible
to use a cause consequence approach. However, these organisational factors can be considered as “indicators” which are symptoms of a future incident. This approach requires attention: the proposed subsets must not be too generic, to avoid a useless classification, nor too specific, to avoid the peculiarities of the particular incident
analysis having an influence on the purpose of the classification. A selection of five recurrent factors is now discussed, taking inspiration from [48]. Some of them may be observed simultaneously, sometimes with mutual enhancement.
The first one is the weakness of the organisational safety culture.
Anticipating what presented in Paragraph 6.5, and quoting [48], the organisational safety culture is intended as a set of factors “put in place or favoured by a business, which concur to achieving the latter's production objectives thanks to the safe functioning of its operation processes.” At the base of this culture there is a sound structure, made of procedures, behaviors, best practise, design
criteria, and so on. Therefore, the safety culture is not the mere sum of the conduct of each actor (who acts prudently and rigorously), but it is reached with a holistic approach. Some of the indicators related to this factor include: managerial deficiencies in safety instructions, the
absence of risk analysis, inappropriate training, and “practices” in conflict with regulations.
A further factor is a complex and inappropriate organisation. The introduction of protection systems in process industries resulted in a safer complex system. This complexity is at the origin of the system's failure, when the organisation (i.e. the system of relations between the different actors) is inadequate. This means that having a large
structured organisation may have the consequence of negatively affecting the decision making process. The indicators include:
coordination problems, lack of each owning responsibility, excessive tasks definition, and poor planning.
The limits of operational feedback are another organizational recurrent factor. Operational feedback is crucial to ensure safety
enhancement. A superficial incident analyses, not taking into account the unfavourable organisational factors related to safety or giving to much space to formalism, is an example of possible “indicator”, as well
as censorship for some aspects of the analysis.
The unfavourable conditions for a safe environment may also be generated from production pressures, together with uncontrolled financial constraints. For instance, a culture that pushes over
production imperatives or a financial approach to safety, considering risks as adjustment variables, are examples of “indicators” for these factors.
The last organisational recurrent factor is the failure of the control organisations. Generally, in order to guarantee an acceptable safety level, risky industries manage their performance with internal audits and monitoring measures. What is questioned is the reliability of these self controls, their effectiveness. From this point of view, some
“indicators” are: the presence of a conflict of interest between
controllers and controlled, the lack of independence in the company, and its tendency to make obstacles for internal formal safety audits and analyses.
In Paragraph 2.3, the “Goal Zero” policy was presented, adopted from many industries to prevent, control and limit the incidents. The above mentioned performance indicators intend to take a picture of the overall safety level of the company at a certain time. They should be monitored to follow how incident rates develop in time, taking into consideration the final goal of the adopted safety policy: having no incident. This desire is actually asymptomatic and frankly unrealistic, therefore it becomes crucial to ask yourself if zero accidents is the right goal. It is evident that no one wants to meet an incident in his/her career, but “zero” is a standard of absolute perfection, while the organisation (made of workers and managers, thus humans) is not absolutely perfect and, unfortunately, failures will occur. Perfection leaves no room for human error, which will inevitably occur.
Moreover, it should be noted that asking workers to have zero
incidents does not explain to them how to have zero incidents [9]. In other words, the adoption of a “zero accidents” goal does not make a safety program. The learning process ensuring a safety improvement is consumed by trial and error. This is why the “Zero” is not the right goal: pushing down the numbers to zero, trying to understand when and where the next accident will occur, is the real goal. And in order to
2.2 foresee the next incident, performance indicators assume a key role.