Reconcile data to add value to

Transcription

1 Reconcile data to add value to your information systems Oscar Rotava, PhD Marcelo Malavazi Sigmafine Users Conference San Francisco, April 26,

2 Agenda Petrobras overview Presentation ti objectives Why to implement mass balance Generating reconciled data Using reconciled data: benefits and profits Perspectives and future work 2

3 Petrobras in the World UK USA Mexico Nova York Houston Colombia Equador Peru Venezuela BRASIL Nigeriai Turkey Iran China Libia Tanzania Singapore Angola Japan Focused in Latin America and Atlantic Basin HEAD OFFICE DOWNSTREAM Bolivia Paraguai Rio de Janeiro Uruguai Argentina MARKETING E&P GAS AND ENERGY AGENCIES Petrobras is an energy company with business activities in exploration, production, refining, trading, transport and distribution of petroleum fuels and energy, in Brazil and abroad. 4 th Energy company in the world (source: PFC Energy jan/2010) 8 th biggest world company by market value: US$ 164,8 billion (source: Ernst & Young - July/2009)

4 Overview of Petrobras (2009) Investments: US$ 40.3 billion Net earnings: US$ billion Net profit: US$ 16.5 billion Shareholders: Crude oil production: 2,562,000 Bbl/d + 413,000 Bbl/d natural gas Production Platforms: 133 Refineries: i 15 Ships: 172 Pipelines: 25,966 km Biofuels: 5 factories(3 in production + 2 in testing) Powerplants: 18 Wind powerplants:1 pilot plant Stations: about 8,000 Fertilizer: 2 factories

5 Refineries Capacity (kbbl/d) 1-Paulínia - Replan Landulpho Alves Rlam 279 Refining Area Overview 3-Duque de Caxias Reduc Henrique Lage - Revap Alberto Pasqualini Refap Pres. Getúlio Vargas Repar Pres. Bernardes RPBC Gabriel Passos Regap Manaus - Reman Capuava - Recap Fortaleza - Lubnor 7 12-Pasadena USA-TX Ricardo Eliçabe Argentina San Lorenzo Argentina Okinawa - Japan 100 Total 2.223

6 Presentation objectives To show the usage of SIGMAFINE in the generation of reconciled data in many different tasks in oil refineries improving i data quality optimizing refinery operation, resulting in profits Hard work and perseverance

7 Why implement a mass balance? Daily mass balance is fundamental in order to establish the accuracy of flow measurements. This will then give good indications of where to make improvements in the management of the HC within the refinery. Economic justification can then be correctly evaluated. Evaluate energy index and oil loss KPI: Determination of fuel consumption and oil loss. Oil is money. IF IT IS NOT CORRECTLY MEASURED, IT CANNOT BE CONTROLLED.

8 Why implement a mass balance? SIGMAFINE in PETROBRAS 10 Refineries in Brazil have SIGMAFINE operational. Not all of them are using it. Mass Balance Workshops /April 2005/March 2007/Junho 2009

9 Generating reconciled data PI Process data Tank transactions Constraints Constraints Sigmafine Reconciled data

10 Using reconciled data Reconciled data can be used in many daily activities of the refinery staff Oil loss accountability Determination of fuel consumption KPI Production accountability support Maintenance support Test run data reconciliation Refinery KPIs follow-up Optimization group support Inventory monitoring Production Planning

11 Using reconciled data: benefits Measurement of Atmospheric Distillation Unit (U200A) feed Definition of the problem Gathering data for analysis Analysis and action Results

12 Using reconciled data: benefits Results Better light crude feed accountability. Calibration results of FC After calibration Before calibration (19/09/2007) Mass flow (t/d) Period Reconcilied FC4051

13 Using reconciled data: benefits Results After calibration the error felt considerably from 7% to 1%. Relative error 8.00% 7.00% Before calibration After calibration (19/09/2007) 6.00% 5.00% Error 4.00% 3.00% 2.00% 1.00% 0.00% Period

14 Using reconciled data: benefits Need Study of distillation unit yields Analyze difference between theoretical and actual yield Why? Yield of units apart from programmed. Impact on refinery KPI Delivery to client impaired i Profitability not optimized

15 Using reconciled data: benefits Tool description Gather composition data of item from terminal and calculate theoretical yields. Calculate reconciled yields. Print report of difference Theoretical yields CENPES Petrofine Item data Actual yields Sigmafine Terminal Refinery

16 Using reconciled data: benefits Analysis made yields of other units

17 Using reconciled data: benefits Lab analysis made Analysis of crude quality API theoretical X API lab 6 5 delta A API Dif. Abs. U-200 Dif. Abs. U-200A Limite

18 Using reconciled data: benefits Results Computational tools used in the development of production planning, tuned with data closer to reality. Difference between database crude yields and processed crude yields above tolerance. Validation of refinery yields. Study about new methodology of yield calculation l when working with blends of crude. Tool increases reliability of accomplishing the production planning.

19 Using reconciled data: profits Nominal feed Above maximum: environmental sanction Below maximum: financial loss $

20 Using reconciled data: profits Measurement analysis: actions Measurement must always be followed by its reconciled value Maintenance report Reconciled measurement < Measurement Lost processing capacity Reconciled measurement > Measurement Check if measurement is inside operational/ environment limits or licensed value In case of doubt program a test with segregated tank.

21 Using reconciled data: profits Nominal feed of hydrotreater U283A Measured X Reconciled values 5800 Carga U283A Loss of processing capacity U$ 25.7M/year Carga Processada (m3) Carga medida Carga Reconciliada

22 Using reconciled data: profits Proof Data-2008 Medidor (m3/d) Varição de nivel (m3/d) Diferença (m3)/d Tanques 03-jan jan mar abr The segregated feed tank showed a surplus difference of 250 m 3 /d due to lower processing feed. The same action in a FCC Unit and a Distillation unit accrued the gain to US$ 33.8M/year

23 Using reconciled data: losses Losses on carbon black feedstock tank because of wrong gpp pipe alignment Identification of the difference In an imbalance of around 500 m³ in the slurry production of FCC units was identified ; Slurry production should be higher according to tank levels; Slurries produced in both FCC Units were aligned to fuel oil tank;

24 Using reconciled data: losses Context analysis FCC annalists informed that because of equipment, feed, and yield constraints, such surplus slurry production was infeasible; Possible wrong alignment of pipelines on the heavy products area: Gasoil to FCC Unit or Vacuum residue; Reconciliation group then began to suspect of non identified gravitational flow between tanks on the heavy products area, where tanks operate simultaneously, in running mode.

25 Using reconciled data: losses Actions taken Reconciliation group waited for a change of operation mode of FCC Units from fuel oil to carbon black feedstock, since in this mode, tanks used are smaller and products are delivered segregated from FCC Units to tanks.

26 Using reconciled data: losses Carbon black feedstock operation mode In carbon black feedstock operation mode, slurry is specified inside units battery limits; The difference persisted from day to ; In day carbon black feedstock was delivered to clients. One specification was outside limits; Possible contamination, ti but contaminant t unknown.

27 Using reconciled data: losses Wrong pipe alignment Reason for the difference: wrong pipe alignment, delivering vacuum residue to carbon black feedstock tank; Flow estimate of the undue flow, around 450 m³; 1, m³ delivered to clients and remaining stock, 2, m³, was reprocessed on FCC Units.

28 Using reconciled data: losses Financial calculations (US$) Carbon black feedstock is a product of higher aggregated value than vacuum residue; Price carbon black feedstock Price vacuum residue Accrued carbon black 1,806, Accrued vacuum residue 1,208, Loss 598,268.55

29 Using reconciled data: losses Intangible losses No compliance with client Reprocessing of finished product Loss of production reliability

30 Perspectives and future work Include loss models into SIGMAFINE model Migration to SIGMAFINE 4

31 Thank you