Accident-Deviation Models

Accident or mishap deviation models as used in system safety processes can permit analysis of events in terms of deviations. The value assigned to a system variable becomes a deviation whenever it falls outside an established norm. When measuring system variables, these deviations can assume different values depending on the situation. Hazard control policies and procedures should detail any specified requirements. A deviation from a specified requirement could result in a human error for failure to follow procedures. Therefore, we must consider incidental factors as deviations from an accepted practice. An unsafe act relates to a personal action that violates or deviates from a commonly accepted safe procedure. Time functions as the basic dimension in a system deviation 

Accident and mishap deviation models used in system safety processes can permit analysis of events in terms of deviations. The value assigned to a system variable becomes a deviation whenever it falls outside an established norm. When measuring system variables, these deviations can assume different values, depending on the situation. Hazard control policies and procedures 

Develop an open-door policy to provide employee access to managers. 

Develop an easy-to-use accident, injury, and hazard reporting system. 

Require supervisors to conduct periodic safety meetings. 

Disseminate hazard control and accident information in a timely fashion. 

---

Explain the basic premise of accident deviation models. 

Certain types of deviations within a man-machine system increase the probability of accidents and/or the expected value of loss due to accidents, thus being valid indicators of the accident risk. The nature of accidents represented by the deviation model is based on systems theory, on the energy exchange model, the multilinear events sequencing approach, and concepts of human errors as discussed in chapter 4. 

The answer to the second question was obtained by using models from organizational control theory. A deviation can re-occur due to ineffective operation of the organization s control process. A theoretical model of this control process, in which causes of precursors can be expressed in terms of ineffective control elements of the organization s control process, was derived from existing models in literature. However, as safety literature shows, there are certain conditions shaping a situation that make these control elements ineffective. These conditions, sometimes called latent conditions in safety literature are the actual root causes of precursors and possible accidents. In this thesis a classification has been developed which identifies six main types of latent conditions (these six latent conditions are context related but... 

The answer to the question is that in the vast majority of all accidents, re-occurring deviations, which were defined as precursors, are present. That a variety of events were present prior to accidents, was stated by Heinrich (Heinrich, 1931). He combined the common cause model and the descriptive iceberg model, stating that prior to an accident, increased numbers of near misses, errors and recoveries are present. Subsequently, Turner (Turner, 1978) identified a so-called incubation period prior to accidents, in which all sorts of events occur unnoticed or are misinterpreted. However, neither study indicates types, or categories of events that can act as precursors of accidents. 

Not just by accident PLS regression is the most used method for multivariate calibration in chemometrics. So, we recommend to start with PLS for single y-variables, using all x-variables, applying CV (leave-one-out for a small number of objects, say for n < 30, 3-7 segments otherwise). The SEPCV (standard deviation of prediction errors obtained from CV) gives a first idea about the relationship between the used x-variables and the modeled y, and hints how to proceed. Great effort should be applied for a reasonable estimation of the prediction performance of calibration models. 

Like HAZOP, STPA works on a model of the system and has guidewords to assist in the analysis, but because in STAMP accidents are seen as resulting from inadequate control, the model used is a functional control diagram rather than a physical component diagram. In addition, the set of guidewoids is based on lack of control rather than physical parameter deviations. While engineering expertise is still required, guidance is provided for the STPA process to provide some assurance of completeness in the analysis. 

As mentioned previously, the analysis of deviations which are related to the overall task goal and appH-cations can be critical. This is supported by the Task Modeller. For example, if the operator omits a step of a procedure and the omission can in some cases lead to an accident, this potential omission needs to be analyzed in sufficient detail to provide the information required by initial scopes of the case study. 

Having made the decision to model hazards at the level where driver behaviour deviates from the ideal, which may not result in an accident, the accident frequency can be broken into two factors, as follows ... 

For this paper we treat hazard assessment as a combination of two interrelated concepts hazard identification, in which the possible hazardous events at the system boundary are discovered, and hazard analysis, in which the likelihood, consequences and severity of the events are determined. The hazard identification process is based on a model of the way in which parts of a system may deviate fi om their intended behaviour. Examples of such analysis include Hazard and Operability Studies (HAZOP, Kletz 1992), Fault Propagation and Transformation Calculus (Wallace 2005), Function Failure Analysis (SAE 1996) and Failure Modes and Effects Analysis (Villemeur 1992). Some analysis approaches start with possible deviations and determine likely undesired outcomes (so-called inductive approaches) while others start with a particular unwanted event and try to determine possible causes (so-called deductive approaches). The overall goal may be safety analysis, to assess the safety of a proposed system (a design, a model or an actual product) or accident analysis, to determine the likely causes of an incident that has occurred. 

It is well known that speed is a crucial road safety factor. Many implemented safety measures aim to induce road users to reduce their speed and comply with speed limits. With respect to the change in the mean speed, the impacts on road safety in terms of number of accidents and the number of injured and killed people are well known. For example, this relationship is described by the Power model [ELV 04, NIL 04], which is often used to estimate the traffic safety effects of speed changes. However, it is not certain that only the mean speed is affected by particular traffic safety measures measures such as the 85th percentile, standard deviation of speed and shape of the speed distribution can also be affected. 

The INRS Model builds on the principles of fault-tree analysis (Leplat, 1978). The model focuses on variations or deviations from the usual course of work at the work-systems level. There are four classes of variations, those related to the individual, the task, equipment and the environment respectively. We here see a clear relation to ergonomic models of work systems. The findings from an accident investigation are displayed in an analytic tree, showing causal relations. Figure 5.9. It gives a schematic presentation... 

Human information-processing models focus on the interaction between the human operator and the environment in a disturbed system. This interaction is analysed from the human operator s point of view. The operator is viewed as an information processor who responds to deviations and hazards in the environment, Figure5.10.Inananalysisofan accident, the aim is to identify human failures in identifying and evaluating the situation and in taking the appropriate measures. 

Incidents rarely occur by pure chance. In the analysis of how incidents occur, we rely on the OARU model of Section 5.3. An incident is usually preceded by deviations at the workplace that increase its frequency and/or the consequences of it. SHE measures that eliminate existing deviations (e.g. repair of faulty safety equipment) will have an immediate effect on the risk of accidents. They will, however, not have lasting effects if the deviation may occur again. 

We will later apply the accident-analysis framework in a review of different types of methods used in the collection and analysis of data of accident risks. We will start at the output side of the model by reviewing the different types of classification systems used to document the consequences of accidents and different measures of loss. We will then continue by looking into the classification systems used to document incidents and deviations. Finally, we will review the different classification systems for contributing factors and root causes. Our aims will be twofold first, to be complete, i.e. by presenting all alternative means of measuring and classification, and second, to give specific advice on the preferred method. The reader will find recommended alternatives in shaded tables and checklists. 

Diagnosis is defined as the complete decision cycle consisting of (1) identification of symptoms, (2) determining causes and (3) prescription of remedy. A symptom is a deviation of the system s behaviour from what is considered to be normal , a term we are familiar with from the process model of accidents (Section 5.3). 

Both the integrative model by Smillie Ayoub (1975) and the deviation concept by Kjellen (1984a) connect the general systems theory to the sequencing and energy models of accident causation. They encompass technical, organizational and human components of the system. Various methods of system safety analysis (e.g. fault tree analysis, incidental factor analysis) support the identification of technical and human deviations as well as the analysis of the conditions and consequences of these deviations. From the discussion of near misses and conflicts it became clear that frameworks of accident causation should cover all kinds of incidents, thus becoming frameworks of incidents. 

The analysis of many accidents has led to the appreciation that multiple equipment failures and process deviations combined with faulty human decisions and actions are often involved. Safety assessments, therefore, are not complete unless the interactions between equipment failures and human actions are considered. Since human behaviour is complex, and does not lend itself immediately to relatively straightforward reliability models, it is suggested that the following classifications of human interactions (that typically group all activities) need to be considered (Mahn et al. (1995)) ... 

---