Causal event sequences

Error analysis techniques can be used in accident analysis to identify the events and contributory factors that led to an accident, to represent this information in a clear and simple manner and to suggest suitable error reduction strategies. This is achieved in practice by identification of the causal event sequence that led to the accident and the analysis of this sequence to identify the root causes of the system malfunction. A discussion of accident analysis techniques is included in Chapter 6. 

Keywords acausal reasoning, case relations, causal accounts, causal event sequences, causal reasoning, constraint-based reasoning, current electricity, device model, dynamic physical model, dynamic processes. Educational Testing Service, electricity, electrostatics, envisioning, macroscopic models, naive physics, physics, prior knowledge, qualitative arguments, qualitative model, qualitative theory, transient processes... 

On the basis that a picture is worth a thousand words, a sequence diagram can be used as an effective tool for commimication about the incident sequence, the initiating events, and their causal factors. Sequence... 

The event and causal factors charts used by the Department of Energy as an accident (or mishap) investigation tool is basically the same tool as multilinear event sequencing (MES) and similar to simultaneous timed event programming (STEP)—all based on concepts of Ludwig Benner. It is also called causal factors analysis (CFA). 

The basic concept from which event and causal factors charts were developed can probably be traced back to Ludwig Benner and others at the National Transportation Safety Board. Benner developed a very similar technique called multilinear event sequencing (MES) and more recently sequentially timed events plotting (STEP). Event and causal factors charts were part of the overall MORT approach to system safety developed by W. G. Johnson for the Atomic Energy Commission in the early 1970s and further developed and taught by the Department of Energy s System Safety Development Center (SSDC). The use of the event and causal factors chart is sometimes referred to as causal factors analysis. 

These are followed by the analysis of the accident and the causes. At the end, the report is concluded by presenting the causal chain of events and the underlying factors in the accident, as well as some recommendations to improve maritime safety. The parts that are fully reviewed for this study are summaries, analyses, and the conclusions. However, for some of the reports, other parts are also browsed in order to better understand the accident and the sequence of the causal events. 

When the diagram is complete, the analyst proceeds through it to identify sets of events that were critical in the accident sequence. These critical events are then subjected to a further causal analysis using other techniques such as root cause coding, described below in Section 6.8.4. 

The events and causal factors chart for this incident is shown in Figure 7.9. The primary sequence of events is shown horizontally in bold boxes. Secondary events are shown in the other boxes, and conditions are in ovals. From the diagram three causal factors were identified and carried forward to the Root Cause Coding to establish the root causes of the causal factors. 

Time. Physical events generally occur in some causal sequence. Time is a measure of this sequence and is required in addition to position in space in order to fully specify an event. 

For acute releases, the fault tree analysis is a convenient tool for organizing the quantitative data needed for model selection and implementation. The fault tree represents a heirarchy of events that precede the release of concern. This heirarchy grows like the branches of a tree as we track back through one cause built upon another (hence the name, "fault tree"). Each level of the tree identifies each antecedent event, and the branches are characterized by probabilities attached to each causal link in the sequence. The model appiications are needed to describe the environmental consequences of each type of impulsive release of pollutants. Thus, combining the probability of each event with its quantitative consequences supplied by the model, one is led to the expected value of ambient concentrations in the environment. This distribution, in turn, can be used to generate a profile of exposure and risk. 

If X is space-like and the events are designated such that t2 > 11, then c(ti — f2) < z — z2, and it is therefore possible to find a velocity v < c such that ic(t[ — t 2) = X vanishes. Physically the vanishing of X means that if the distance between two events is space-like, then one can always find a Lorentz system in which the two events have the same time coordinate in the selected frame. On the other hand, for time-like separations between events one cannot find a Lorentz transformation that will make them simultaneous, or change the order of the time sequence of the two events. The concepts "future" and "past" are invariant and causality is preserved. That the sequence of events with space-like separations can be reversed does not violate causality. As an example it is noted that no influence eminating from earth can affect an object one light-year away within the next year. 

The important point in the present context is that these cognitive abilities do not come for free. It is clear that high levels of intensionality are extremely difficult to cope with in computational terms. Kinderman et al. (in press), for example, tested normal adults with a series of tests similar to those used in standard ToM tests but which allowed for up to fifth order intensionality (as opposed to the conventional second order of standard ToM false belief tests). At the same time, subjects were also given tests of environmental causal relationships that required only memory of a sequence of events. Memory tests involved causal relationships of up to sixth orders of embeddness ( A caused B which caused C which. caused F ). Error rates on memory tasks varied fairly uniformly between 5-15% across the six levels of embeddness with no significant trends in contrast, error rates on the ToM tasks increased exponentially with order of embeddness (i.e. intensionality). 

A component for identifying the critical events and conditions (causal factors) in the incident sequence. 

The timeline tool pulls all of this information together into a manageable record of events and sequence providing a perspective conducive to proper causal analysis. 

Sequence diagrams are a more elaborate graphical depiction of a timeline, and allow the investigator to present related events and conditions in parallel branches. These sequence diagrams are also known as causal factor charts. 

Events Causal Eactor Charting (E CE) (5) was adopted by the developers of MORT to identify and document the sequence of events leading to an incident. A number of proprietary process safety incident investigation methodologies, such as SOURCE ) and TapRooT C) include E CF as one of their building blocks. 

The first step in developing a causal factor chart is to define the end of the incident sequence. Construction of the chart should start early from the end point and work backward to reconstruct what happened before the incident by identifying the most immediate contributing events. 

Find the facts in the main sequence on the Causal Factor Chart that describe a component failure or a human error. Ensure the fact is not describing a management system failure (i.e., ensure the fact is not a root cause, near root cause, or root cause category). The identified negative events/conditions are candidate causal factors. Any candidate causal factor that is not dependent on another candidate causal factor is a valid causal factor. 

The next phase of investigation involves developing a preliminary chronological description of the sequence of events that led to the failure. Timelines can be developed in various formats and levels of detail, from simple lists of events to complex sequence diagrams or causal factor charts, usually dependent upon the particular circumstances of the investigation being conducted. 

Once the evidence has heen collected, a timeline or sequence diagram developed, and the actual scenario confirmed, the investigation can proceed to the next stage, the identification of causal factors. These causal factors are the negative events and actions that made a major contrihution to the incident. 

The design of most process plants relies on redundant safety features or layers of protection, such that multiple layers must fail before a serious incident occurs. Barrier analysis ) (also called Hazard-Barrier-Target Analysis, HBTA) can assist the identification of causal factors by identifying which safety feature(s) failed to function as desired and allowed the sequence of events to occur. These safety features or barriers are anything that is used to protect a system or person from a hazard including both physical and administrative layers of protection. The concepts of the hazard-barrier-target theory of incident causation are encompassed in this tool. (See Chapter 3.)... 

There are a number of quality assurance checks that should be considered before identifying the final list of causal factors. It is important to test for sufficiency of the information when compiling a sequence diagram. This test for sufficiency may be performed by asking one or more of the following questions, when comparing two adjoining facts in the sequence of events ... 

Thus, a deterministic answer assumes that the laws of physics and chemistry have causally and sequentially determined the obligatory series of events leading from inanimate matter to life - that each step is causally linked to the previous one and to the next one by the laws of nature. In principle, in a strictly deterministic situation, the state of a system at any point in time determines the future behavior of the system - with no random influences. In contrast, in a non-deterministic or stochastic system it is not generally possible to predict the future behavior exactly and instead of a linear causal pathway the sequence of steps may be determined by the set of parameters operating at each step. 

---