Getting Value Out Of Reconciled Data Using PI and Sigmafine Ales Soudek Business Development
Outline What is the PI System? What is Reconciled Data/Sigmafine? Using PI to Manage Your Assets What Value to the Business? Examples from Companies
The PI System High Availability Context will apply to multiple PI servers Analysis ACE PIANO Sigmafine Analysis Tools will work against a single or multiple PI Servers PI System 1 PI System Interfaces 2 Server Functions Time-Series Interfaces Batch PI System N Time-Series Server Functions Batch Time-Series Interfaces Batch Time-Series Server Functions Batch Standards Standards General PIBAGEN PIBAGEN Event File Snapshot Snapshot Archive History Time-Series Batch Time-Series Batch History Search Standards PIBAGEN Snapshot History Com General Event File Archive Search SQC Connector Com General Event File Archive Search SQC Manual Connector Alarms Com SQC Manual Connector Alarms Enterprise Equations Manual Alarms Enterprise Equations Data Access Data Tools Access Data Tools Access PI-SDK Tools PI-SDK PI-API PI-SDK PI-API PI-OLEDB PI-API PI-OLEDB PI-OLEDB Context Models Analysis Framework Clients General ProcessBook DataLink RtReports WebParts ActiveView Specialized AlarmView PI-SQC BatchView RtAlerts Enterprise Equations Asset Model with features from AF and MDB Standardized interfaces for assets and non-time-series data Foreign Assets S95 MIMOSA Foreign Data RDBMS WS
Reconciled Data PI Manual Data Process Data Movements Tank Inventories Analysis Gross Error Check Model Framework Sigmafine Reconciliation Reconciled Data
Sigmafine Analysis Principles Balance Gross Error Detection Weighted Least Squares Data Reconciliation Uncertainty := (Bias 2 + Precision 2 ) 0.5 Fn : = N Measi Recon i = Toli 1 i 2.00-10.00-20.00-30.00-40.00-50.00-60.00-70.00-80.00-90.00 Meter 18 y = 0.7732x - 13.639-100.00 meter18 Linear (meter18).00
The Reconciliation Process Model = Plant? Daily Mass Mass Balance 100%?? YES NO Gross Error Detection Gross Error Correction Continuous Improvement Daily Data Reconciliation YES NO CHANGE Official Production Information
The Data Fog Need for better information and faster access Flood of data Poor data quality - loss of confidence in reported results Data does not balance Manipulation by different groups Direct financial impact Custody transfer errors go undetected Plant operation is sub-optimal Data Info Fog?
Lifting the Data Fog - Sigmafine Validated data for ERP Standardization One Set of Numbers Detect Measurement Problems Better Business Decision Data Info Fog
Templates of Assets These assets can be made into a template for reuse. Asset Attributes: Name plate information Trays Related data references Limits Asset Attributes: Name plate information Related data references Limits Asset Attributes: Name plate information Related data references Limits Asset Attributes: Name plate information Related data references Operating limits Asset Attributes: Name plate information Related data references Maintenance information Tower 1
Use and Manage Assets Collection Based Analyses Plant 1 Equipment Based Analyses Tower 1 Connectivity Based Analyses η = 90% η = 65% Mass Balance Energy Balance Component Balance η = 74% η = 94% Tower 1
PI in the Business Cycle Running plan definition PI-AF PI-PB Assign plan to operations Cockpits RtWebParts PI-Datalink PI-PB PIACE Automatic procedures Reports Analysis PIACE PI-AF Sigmafine KPI calculations Consistency through OSIsoft infrastructure Modeling & Reconciliation
Return on Investment Keep Focus on Target Objectives Rapid Identification of Yield Degradation Historical Analysis Redefine Targets for Improved Performance Increased Diesel Yield 2.2% on Feed $1,200,000
Organizational Value Reduction in Deviation Between Planned vs Actual Yields (e.g. diesel from 2.2% to 1.3%) Demonstrated how sharing information and involving people helped to reach the refinery targets PI System allowed engineers to focus on the real business objectives Performance management approach allowed to keep production and plant management under control with a lean organization
Reconciled Data Value Monitor and Reduce Meter 12 Maintenance % Bad Meters 10 8 6 4 2 0 May Jul Sep Nov Jan Mar May Jul Measurement Validation Drift Bias Pure Gas E 1750 1600 1400 1200 1000 800 600 400 C0F1230.AD 1649.4 C0F1230.SF B2F1003.AD 400.5 B2F1003.SF B2F2003.AD 402.47 B2F2003.SF F0F1013B.AD 19.939 F0F1013B.SF Pure Gas W 1750 1600 1400 1200 1000 800 600 400 20F0101.AD 1571.4 20F0101.SF 12F1003.AD 422.08 12F1003.SF 12F2003.AD 428.68 12F2003.SF F0F1013A.AD 0. F0F1013A.SF 200 200 0 02-Sep-01 00:00:00 100.59 Day(s) 11-Dec-01 14:02:54 0 11-Nov-01 00:00:00 30.59 Day(s) 11-Dec-01 14:02:56 Meter Error SLO East 250 200 150 100 50 0 02-Sep-01 00:00:00 100.59 Day(s) 11-Dec-01 14:03:04 C0F1109M.AD 89.787 ton/h C0F1109M.SF ton/h C0F1116M.AD 76.712 ton/h C0F1116M.SF ton/h C9F1004.AD 181.66 m3/h C9F1004.SF m3/h SLO West 250 200 150 100 50 0 02-Sep-01 00:00:00 100.59 Day(s) 11-Dec-01 14:03:05 20F1109.AD 99.745 m3n/h 20F1109.SF m3/h 20F1116.AD 89.789 m3/h 20F1116.SF m3/h 29F1004.AD 192.7 m3/h 29F1004.SF m3/h
Tracking Data Quality
Reconciled Data Value Mass Balance before Sigmafine Feb/01 - Jan/02 (Loss % Cumulative = 1.09%) 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% -6.0% (%) 5 4 3 2 Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan 1 0-1 Loss % 3 Month Rolling Loss % Cumulative -2-3 jan-04 mar-04 mai-04 jul-04 sep-04 nov-04 jan-05 mar-05 mai-05 Loss Accountability jul-05 sep-05 nov-05 Gross error detection Regular improvements Sigmafine available Optimization SF Losses jan-06 mar-06 mai-06 jul-06 Loss Accountability Oil loss figures after Sigmafine implementation Benchmark 1.08 0.23 Solution 0.12 sep-06 nov-06 jan-07 0.48 mar-07 mai-07 0.30 0.53 Jun/02 Jul/02 Aug/02 Sep/02 Oct/02 Nov/02 SF montly SF cumulative 1.20 1.00 0.80 0.60 0.40 0.20 0.00 %
Reconciled Data Value Mass balance yield - FCC (Fluidized Catalitic Cracking Unit) Mass balance yield (%) 110 105 100 95 90 85 80 *Data reconciliation group identified LCO flow meter fault *During this period it was believed that LCO yield had increased. Eliminated loss corresponding to US$ 23.5 million per year Unit shut down Different catalyst loaded 28-01-2005 17-02-2005 09-03-2005 29-03-2005 18-04-2005 08-05-2005 28-05-2005 17-06-2005 07-07-2005 27-07-2005 Date Flow meters maintenance required by reconciliation group **LCO flow meter repaired **Actual LCO yield equal to the value previously to the catalyst change **Lower light HC and more decanted oil produced = Less profits
Conclusions Monitor and reduce meter maintenance Measurement stewardship Verify custody transfer Monitor plant performance Identify operations problems Loss identification and tracking Business Decisions Based on Better Data
Trivia History Lesson
Computer of the Future (1954)
Computer of the Future