THE DEVELOPMENT OF MODELLING FOR PHYSICAL PROTECTION SYSTEM EFFECTIVENESS EVALUATION AT NUCLEAR FACILITIES UNDER UNCERTAINTY

• Main Author: Andiwijayakusuma, Dinan
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75717
• Institution: Institut Teknologi Bandung

The Physical Protection System (PPS) integrates various protection elements of personnel, procedures, and devices to protect assets, materials, and facilities from theft, sabotage, or other attacks. The main functions of the PPS are (a) deterrence: which prevents attacks by providing a caution to prevent adversary intrusions (b) detect: detects any adversary intrusions (c) delay: delaying adversary movements after the intrusion occurs and is detected; and (d) response: response is the reaction carried out by a response force unit to interrupt and counteract the adversary. In its application, we must evaluate the effectiveness of PPS periodically. The evaluation aims to ensure that PPS still meets the requirements of the PPS design objectives to achieve a reliable PPS that aligns with the dynamics and technological developments that can affect the effectiveness of PPS. Evaluation of the effectiveness of the PPS can be done through two methods: the compliance method and the performance method. Evaluation of the PPS with a compliance method is to evaluate the effectiveness of the PPS based on the suitability of the facility with specific compliance requirements and regulatory policies. Evaluation of PPS with a performance approach evaluates each protection element and its effect on overall system. The advantage of this analysis is that it supplies the qualitative and quantitative analysis techniques, with an essential aspect of quantitative analysis using numeric elements for PPS. The modeling development for the evaluation effectiveness of PPS in this study uses a performance approach based on EASI (Estimated Adversary Sequence Interruption). EASI is a single-path analysis tool to calculate the probability of interruption (!$) to find the most vulnerable path in a facility. EASI is popularly used and has many advantages, but there are limitations in evaluating PPS. EASI uses single-path analysis, so a PPS analyst must determine all the adversary's paths and then analyze them one by one. EASI can potentially problematize the PPS analyst in analyzing a larger facility with more extensive possible pathways. This research proposes a multi-path model to overcome this problem. Another limitation of the EASI model is that the calculation results only produce a single value of the probability of interruption (!$) without the associated uncertainty. It occurs because the EASI model uses a single value only for each protection element's detection and communication parameters without considering the uncertainty in detection performance. This research develops a model that produces the distribution for the value of detection and communication parameters. The method uses a stochastic approach through Monte Carlo simulation. The probability of detection (!%) and the probability of communication (!&) data sampling was carried out for probability of interruption (!$) calculations involving uncertainty quantification. The development of this model can also determine the most vulnerable path used by the adversary (Most Vulnerable Path - MVP) by implementing the concept of a critical detection point (Critical Detection Point - CDP). We test the new model using a hypothetical facility (Hypothetical National Nuclear Research Facility - HNNRF) to estimate the PI value. The calculation starts by converting a hypothetical two-dimensional (2-D) facility schematic into an Adversary Sequence Diagram (ASD). We can search the CDP through the ASD and determines the most vulnerable path (MVP). The calculation results show that the PPS effectiveness for HNNRF is 80,3%. Using variability and uncertainty of !% and !& values for the calculation of !$, the results show that the distribution of !$ values illustrates the potential for a significant reduction in the estimated value of !$ under certain conditions. It should be an essential concern for designers and SPF analysts to determine the !% value correctly or set a higher standard for reliable SPF performance. The sensitivity analysis of the estimated !$ value shows that the relationship between !% to !$ and !& to !$ is linear. The sensitivity analysis shows that an increase in the probability of detection of protection devices located on the facility's outer layer significantly increases the !$ value compared to the protection elements in the layer closest to the target.