Universidade Estácio de Sá Engenharia de Produção MODELING AND SIMULATION OF PIPELINE LOADING OPERATIONS ONTO BARGES: A CASE STUDY OF RESOURCE ALLOCATION USING DISCRETE EVENT SIMULATION Fabricio Cardoso de Vasconcellos Flávia Cristina da Silva Duarte Henrique Alves Serpa Prof. Dr. Marcelo Prado Sucena Prof. Dr. David Fernandes Cruz Moura
Agenda Introduction Loading Operations Characterization; Discrete Event Simulation Architecture; Case Study Construction Steps Verification & Validation; Technical Scenario Analysis (What-If); Economic Scenario Analysis; Final Remarks.
Introduction Problem: Reduction of Loading Time of Pipelines onto Barges Methodology in Brief: Discrete Event Simulation (DES)-driven resource allocation (trucks, reach stackers, and cranes) analysis of different investment scenarios.
Motivation Several bottlenecks in a Brazilian pre-salt area surronding port (São Sebastião): Berthing average utilization rate: 18 hours 50% higher than optimal values presented in literature Queue generation at the berth area Almost 10% of the overall containers freight costs Loss of port calls
Introduction Source: Carvalho, 2011
Technique Choice Reasoning DES Advantages New process configurations verification Design of novel operational proceedings System evaluation for different timing conditions Easy process bottlenecks identification Model reproducibility Low cost investment
Loading Operations Characterization
DES Project in Brief Source: Chwif; Medina 2006
ACD Conceptual Model
Case Study Steps Reach Stacker Loading Time Intervals Weibull (Fixo= 1, α= 3.48, β= 0.736); Truck traveling time intervals Weibull (Fixo= 2, α= 15.2, β= 1.11); Truck weighting time intervals Pearson 5 (Fixo= 1, α= 3.48, β= 0.736); Crane loading time intervals Pearson 5 (Fixo= 5, α= 5.53, β= 4).
Computational Model Construction Software Simul8 Scenario Representation: Actual logical sequence of pipes (B, C, D, A); Actual proportion of plain pipes (93%) and anode pipes (7%); Attendance of 4 pipes at a given time on each resource(rs, trucks, cranes, and port scale); Each load: 708 pipelines @ barge; Mean Loading Time: 18 hour @ load.
Validation Issues Simulation Model: Pipe Loading Mean Time:18.075,87 min = 301,25 h to load 12.313 pipes. Number of Loaded pipes: I.C (95%) = [12.242, 12.321] At a given load out operation: Actual System: Time average = 18 hours Number of loaded pipes = 12.313 t(hours) = 708 * 301,25 = 17,32 h 12.313
Scenario Analysis Loading Time X # of Cranes 20000 18000 18595 16000 Tempo de operação (min) 14000 12000 10000 8000 6000 9352 7891 7854 7848 4000 2000 0 Número de guindastes (unidades) Fonte: Próprio autor
Optimal Number of Cranes x Loading Time
Scenario Analysis Queue Time & Utilization @ Cranes 1.2 99 % 1 0.8 78 % 0.6 0.4 50 % 43 % 0.2 14 % 0 Cenário 1 Cenário 2 Cenário 3 0 % Tempo % Utilização Guindaste Tempo de Fila Guindaste
Economic Analysis Equipment Rates Scenario 1 Scenario 2 Scenario 3 Trucks (unities) 4 4 4 Cranes (unities) 1 2 3 Total loading time (h) 302 152 129 Cranes rental (US$) 2,700 135,900.00 136,800.00 174,150.00 Reach Stacker rental (US$) Barge berthing tax (US$) Tug berthing tax (US$) Storage yard rental (US$) Manpower (US$) Trucks rental (US$) Total (US$) 2,700 249.50 53.00 753.50 13,840.50 490.00 135,900.00 12,551.87 2,668.67 37,919.87 696,634.47 98,653.33 1,120,228.23 68,400.00 6,317.00 1,343.17 19,085.50 350,623.97 49,653.33 632,223.48 58,050.00 5,361.56 1,139.93 16,197.56 297,569.03 42,140.00 594,608.08 Scenario 2: 44% reduction when compared to #1 Scenario 3: 47% reduction when compared to #1
Final Remarks Investigation of the resource allocation issue such as trucks, cranes and reach stackers at the Port of São Sebastião, to propose a reduction in total loading time of pipelines on barges. Verification and Validation of an actual port model What-if scenario analysis to enhance productivity and suggest improvements in resource allocation. Brief economic analysis of the suggested scenarios proposed by the simulation model Conclusion: Scenario 3 - better performance, but comprises space reduction for the safe movement of equipments, people and products. Scenario 2, therefore, constitutes the best option, as it showed a 50% reduction in the total loading time and cost reduction of approximately 44%.
Final Remarks We conclude that performance evaluation by means of a discrete event simulation methodology allowed the assessment of alternative investment scenarios, constituting itself as a fundamental tool for the characterization of a port terminal of pipeline loading, diagnosing problems and identifying possible improvement opportunities.