(81s) Dependent Layer of Operation Decision Analyzes (LODA) to Calculate Human Factor, a Simulated Case with GLP Event

INTRODUCTION, LOPA. The opportunities of the accident may be in the structures of equipment, methods of management, technology with aspects of process control, the handling characteristics of the product, products application, the opportunities caused by human factors and other factors connected with the project or organizational environment. LOPA is a simplified way of defining risk analysis for each potential safety functions (SIF) and levels (SIL), or is related standards in the field of safety instrumentation for the process industry. This technique evaluates the effectiveness of independent layers of protection in reducing the severity of an undesirable event to meet organizational needs. The technique enables LOPA analyze the investment needed for Safety Instrumented Systems. The vision of layers allows independent calculation of the probability of events, but provides a framework oversimplified does not always happen on the anatomy of an accident (the independence between events). When the hazardous event has simplified training is evident calculate the likelihood and based on type of product released and the consequences on the community, one can calculate the investment protection to avoid shock. The entrepreneur has designed what level of maximum risk to be protected internally and what level is hired externally. In LOPA, the Entrepreneur defines what your intention in terms of risk (business), and the layers are analyzed according to the original design and engineering after the departure of the operation there is a need to correct the risk with instrumented systems as indicated. And about human factors? How to quantify? It is more reasonable to quantify the human factor in risk analysis when it can perform dynamic accident involving dividing into two stages: (1) analysis of the layers of operational decision and (2) analysis of the anatomy of failure (the tree) for measuring and proposition barriers (technical, managerial and operational). LODA. It is known that the operation team behavior moments in differentiate forms between routine and emergency situations or extreme stress (ex: emergency stop). Thus, it is proposed to analyze the behavior of operators in terms of risk, seeking responses from the inventory of knowledge and based on teamwork. This methodology is then named as LODA (Dependent Layers of Operation Decision Assessment) that provides a dynamic means of achieving this state of stress and to examine whether the knowledge inventory in the team are sufficient to address the emergency. In the techniques of dynamic fault tree and operational decisions are taken to prevent states of abnormal move, these events are sequenced at different levels as shown in Figure. In each series event decisions are made by the transaction for signs of failure, as the corrections to the process, on ways to operate or perform tasks and maintain the equipment. The first technique aims to gauge the reaction team and the quality of proposed actions in order to prevent the continuity of the accident. The second technique aims to estimate the frequency of the accident based on the frequency of the events estimated in the first technique, to examine the consequences of the accident for the company and the community, proposing the installation of barriers to prevent the accident at the management level and with safety devices facilities and recalculation of the frequency of accidents. This technique is being applied in the logistics and safety activity in transportation and storage of liquefied petroleum gas (LPG), consisting of a mixture of propane and butane. This activity logistics and warehousing is located in an urban area of a city of the Northeast (Brazil), close to residential properties and other industrial facilities. Although the risks of accidents involving the community are rare, this event is possible and requires preventative screening, including those aspects regarding the identification of latent failure, as indicated in chain abnormalities. The events in the sections for drying and regeneration were chosen to illustrate a simulated explosive situation, since the signs of abnormalities, such as the absence of flame, until corrective measures to reduce the strength of the fault, then came the blast furnace and equipment for "domino effect". These events were based on the speech of the operator (occurrence book) and on assumptions about the continuation of abnormalities. The events in chronological order involved in this risk analysis are listed: (D1) fluctuation of temperature, pump ranges, channel swings, unstable flow regeneration, photocell dirty, solenoid valve failure, flame goes out without indication; (D2) in charge pump, temporary stop at the pump, difficulty level control, momentary stop of flow, N2 purge, gas detector fails, without detecting excess propane, propane with O2, propane with O2 in the oven, cleaned only photocell, off duty , inattentive operator, improper position, overheating, (D3) serpentine weakened, local overheating, leakage of propane in the furnace, passing out of the oven, calibrated detector, propane to the atmosphere without notice, (D4) of O2 in the limit, with torch shows O2 excess noise exceeds five minutes warming up, breaking the coil, dumper restricted blast furnace, the hot spark of service, short conduit, presence of ignition, external fire, failure in taking action; (D5) difference in pressure oven for pots, fire training channel, PSV failed by speed, explosion vessel, fire spreading in pool, high wind speed, calls for community fire affects a ball, bleve, warning ineffective leaving 50% of people at home, people do not realize the explosion, the possibility of burning people, probability of death in quantity. CHRONOLOGY AND STRENGTH. This risk analysis is accompanied by the simulation of measurement of force (energy) where the fault for each operational factor of failure in the task are estimated and entered the time of event and its estimated strength. Measuring the strength of the failure is relative and is related to approach the maximum point of failure in the task. Thus, the maximum force for failure is indicated when some continuous variables and counts are approaching the state of total disarray. This research indicates the chronology of the times, the operational factors that increase or decrease the strength of the fault, and even the shape of failure over time. This exercise is performed, simulating a prepared based on events measured with the staff of the company, without joint occurrence of factors, and with the times estimated by a specialist. DYNAMICS UNDER STRESS. Team decisions. The operation team was present and participated in the Workshop on Dynamic simulation of blast furnace drying in two stages. Five teams with the technical results and behavioral results, helping to define the type of human behavior in emergency situations. Participants were receptive to the dynamic Workshop and integration were high. Responded to all surveys, identification of technical culture and participating in the dynamics effectively. There was an opening to tell about events that happened in the past, that is, security has been transmitted for that to happen. The turnout was small in quantity, but with the effective participation of all - have been clarified aspects of human reliability, its role in society and in the operation. The participants spoke on the topic of the workshop, which demonstrated the interest of participants and clarification on this. Resistance was not an explicit component of leadership and presentation "inhibitory" in one case. In the dynamic environment is a note yet "warm" in terms of stress, calling attention to matters in the area of commitment and competence creation. FAULT TREE. In this fault tree are included concepts LOPA, except the layers that can't be independent. Also used the speeches of the operation and staff to assemble the map of abnormal events (MEA). The analysis was done of the fault tree, where each event was estimated probability, and then with the installation of defense mechanisms, new calculations were performed. Were also used key words, similar to HAZOP for understanding the risks (of formation channel of fire, explosion vessel, scattering the fire pond, flame toward the community, affect the chance of fire ball, burn people and cause death in accident ). The steps indicated in the procedure were performed and fault tree was constructed by decision level of the operation. The scene of the accident with the physical location of the equipment involved and the direction of drag from the fire, the location of the control room and the community is studied in the risk analysis. Key-words: Operational Dependent Decisions; Risk Assessment; GLP processing; Human factor; Failure Tree Analysis