SIGNAL + DRAHT (104) 10/2012 ETCS n 39 Residual risk of overpassing danger points in ETCS level 1 for the conventional network Abdelouahab Attou A conventional network equipped with the interoperable ETCS level 1 system integrated into the existing lateral signalling (mixed level) is considered. This article presents an approach for the probabilistic and explicit evaluation of the residual risk of a train collision due to overpassing of the danger point. The model has identified a constraint between relevant ETCS system parameters such as Release Speed and the maximum speed in shunting and staff responsible modes for achieving a safety target. Consideration is given to mixed passenger and freight traffic running on the infrastructure. 1 Introduction Among the operating incidents that can occur on rail infrastructure, passing a closed signal at danger and subsequent failure to respect the danger point are those that can cause the most serious accidents. It is therefore essential to evaluate this risk accurately and take effective measures. When the movement of trains is supervised by ETCS level 1, reducing the risk within tolerable limits involves setting of essential system parameters such as Release Speed and maximum speed in shunting and staff responsible modes. The choice of these parameters also involves other considerations such as driving comfort (ergonomics) and the train’s ability to run at low speeds. 2 Existing lineside signalling and warning systems situation Implementing a probabilistic model requires data to be collected beforehand in order to evaluate the frequency of occurrence of the hazard and its characteristics. Historical occurrences of overpassing signal at danger are used to evaluate an annual average frequency value. The analysis of overruns also provides statistics on the distances by which the stop signal has been exceeded. The data we have examined reveal two superimposed distributions whose respective average values are one order of magnitude apart (Figure 1). The first bell represents about two thirds of signal overruns. Its average distance is about 40 m, corresponding to the phenomenon of the driver being distracted and thus inappropriate braking. The other bell includes the remaining third of signal overruns. The average here is about 540 m, corresponding to the phenomenon of driver forgetfulness, so that he begins braking at the last moment, i. e. when the signal becomes visible, or after realising that he has overrun. The average value and standard deviation for the estimated overrun distance in the analysed sample are 205 m and 35 m respectively. Reaching the danger point does not necessarily lead to an accident. Other external conditions relating to the occurrence of a trigger event must be satisfied. The probability of occurrence of these conditions is reflected by a statistical risk exposure factor P exp which is of the order of 10–3 for catastrophic collisions. 3 Hazard in ERTMS/ETCS LEVEL 1 3.1 Release speed monitoring The hazardous situation observed in cab signalling (full supervision or on sight mode) is inherent to the Release Speed ( RS) supervision mechanism introduced for operational reasons. As end of authority supervision applies to the “Max Safe Front End” of the train, the release speed is introduced to the extent that the calculation of the braking curve takes into account factors such as the imprecision in the train’s location, causing the train to be immobilised in rear of the end of authority. The release speed allows the train to approach the end of authority at low speed. One choice is to provide the release speed parameter in the Movement Authority transmitted from the trackside. The other alternative, having the release speed calculated by the onboard computer, was not chosen for operational reasons. Given the existence of certain types of freight trains with weak braking performance and certain configurations of the existing infrastructure in which Figure 1: Distribution of the overpassing distance
RkJQdWJsaXNoZXIy MjY3NTk=