Risk Management in a LPG Tank Yard

LUCRECIA APARICIO; ADRIANA BRANDOLIN; STELLA TONELLI

Conferencia; 2nd CCPS Latin American Process Safety Conference; 2010

In the last years, Quantitative Risk Analysis (QRA) has proved to be a versatile methodology to manage chemical risks for the industry as well as government authorities. The basis of QRA is to identify scenarios and evaluate risk by defining the probability of failure, the different consequences and their potential impact (CCPS, 2000). Then, the risk is defined as a varying complexity function of consequence and probability or frequency.  QRA for off-site risk estimation is focused on identifying and evaluating high impact scenarios, especially those that could affect the surrounding population. Regarding risk management inside plant battery limits, the examination of high frequency scenarios becomes more relevant, even though their consequences are limited.  The information about each scenario helps safety professionals, engineers and managers to evaluate and propose effective risk reduction strategies. It can be used for pointing undetected hazards, performing layout modifications, defining ranking for cost-effective inversions, developing on-site and off-site emergency plans, etc..  Usually QRA results are represented by individual risk contours around industrial facilities or individual risk profiles as a function of distance from the risk source. In these approaches, the individual contribution of a particular scenario to the total risk cannot be identified. Moreover, from a decision making point of view, the consequences of the same scenario could be more or less severe depending on its relative position to plant battery limits.  In this work an example of risk management in a LPG tank yard is presented, based in a new representation of the consequence-frequency relation for each accidental scenario. Consequence analysis is performed using Effects (TNO, 2003), a commercial simulation package which implements rigorous models for fire, explosion and dispersion events and  frequency estimations are carried out by means of Risk-A (PLAPIQUI, 2006). The cases considered are located in a consequence-frequency map where consequences are evaluated according to how much they exceed the plant battery limits. The results allow a simple identification of potential scenarios with high off-site impact as well as incidents with high frequency of occurrence.