Stepladder Models

The GEMS model is based on a more detailed model of human performance known as the stepladder model developed by Rasmussen (see Rasmussen 1986) and illustrated in Figure 2.7. In this model, Rasmussen depicted the various stages that a worker could go through when handling a process disturbance. 

The dotted lines in the diagram indicate the various feedback paths that exist to enable the individual to identify if a particular stage of the processing chain was executed correctly. Thus, if the operating team had planned a strategy to handle a complex plant problem, they would eventually obtain feedback with regard to whether or not the plan was successful. Similar feedback loops exist at the rule and skill-based levels, and indicate opportunities for error correction. The application of the stepladder model to a process industry example is given in Appendix 2A at the end of this chapter. 

APPENDIX 2A PROCESS PLANT EXAMPLE OF THE STEPLADDER MODEL... 

In order to explain each box in the stepladder model shown in Figure 2.7 (reprinted on the facing page), we shall use the same batch processing example as in Section 2.6.3. 

It can be seen that the various boxes in the flowchart can be associated with different stages of the stepladder model. For example, the first box on the left corresponds to skill-based behavior and its associated internal failure mechanisms. The second box illustrates the situation (Stereotype Fixation) where the worker erroneously does not change to a rule-based mode when encountering an unusual situation in the skill-based mode (see also the discussion of the GEMS model in Section 2.6.3). 

This is the actual mental function required by the task that failed (see Figure 2.17). In the case study imder consideration the failure was at the Execute Action stage of the stepladder model, since the worker intended to operate the valve for reactor A, so there was no question of failure in the selection of actions. The connection with the task characteristics box indicates the fact that action is a fimction required by the task. 

This method is based on the Rasmussen stepladder model described in Chapter 2. It was first described in Embrey (1986). The basic units of CADET are the critical actions or decisions (CADs) that need to be made by the operator usually in response to some developing abnormal state of the plant. A CAD is defined in terms of its consequences. If a CAD fails, it will have a significant effect on safety, produchon or availability. 

An extension of the tree of causes, called variation diagrams (Leplat and Rasmussen, 1984) was developed to answer some of these criticisms. In this method, the Rasmussen stepladder model of human error (see Chapter 2) is applied to analyze causal factors at each node of the tree. A detailed example of the use of this technique is provided in Chapter 7 (Case Study 1). 

Figure 7. Stepladder model log(kJka ) vs. log (S/D) cascading plots for CftFe bath gas. Calculated curves have been mown for several values. The data points represent experimentally measured values.

Figures 10 through 12 show that none of the tested models successfully fits the C2H6 data over the complete (S/D) ratio range for any single value. The 6-kcal mol" stepladder model exhibits good consistency with the 0.006 < (S/D) < 0.4 mid-range fallofiF results. The...

Whereas the relative variations of values for various collision partners are fairly certain, the absolute values can only be derived as reliably as the calculated k( ) used in the interpretation (see below). This disadvantage, as compared to low-pressure themud activation experiments, is compensated for by the possibility of producing higher excitation. With the high precisions obtainable today, some information on the functional form of the collisional transition probability has also been derived. Figure IS shows exponential and stepladder models for P(E IE)... 

---